Hydrothermal Transformations of Glycerol
into value-added Chemicals using Zeolite-based
Catalysts
Suna Srisamai
Supervised by Prof Klaus Hellgardt
A thesis submitted for the Degree of Doctor of Philosophy
at the Department of Chemical Engineering
Imperial College London UK
(January 2016)
ii
To my dear mother Nium
and in the memory of my dad Kaewta Srisamai
iii
Declaration
This Thesis is submitted to Imperial College London for the degree of Doctor of Philosophy
It is a record of the research carried out between February 2012 and January 2016 by the author
under the supervision of Professor Klaus Hellgardt It is believed to be wholly original except
where due acknowledgement is made and has not been submitted for any previous degree at
this or any other universities
Suna Srisamai
London January 2016
iv
The copyright of this thesis rests with the author and is made available under a Creative
Commons Attribution Non-Commercial No Derivatives licence Researchers are free to copy
distribute or transmit the thesis on the condition that they attribute it that they do not use it for
commercial purposes and that they do not alter transform or build upon it For any reuse or
redistribution researchers must make clear to others the licence terms of this work
v
Abstract
High availability and low price of crude glycerol the main by-product of the biodiesel
industry make it an attractive feedstock for transformations into value-added chemicals The
aim of this thesis was to improve our understanding of the hydrothermal conversion of glycerol
facilitated by zeolite-based catalysts
A range of Lewis acidic Ce- La- Sn- and Zn-doped ZSM-5 and Beta zeolites were
prepared by solid state ion exchange reaction The effect of those catalysts as well as their
parent NH4- and H-forms on the dehydration of glycerol was investigated under hydrothermal
conditions (270-360 degC 55-186 bar 5-300 min) in batch tubular reactors Several reaction
products were detected of which acrolein was the main liquid product with the highest
selectivity of ~38 mol achieved within the first 5 min at 330 degC on H-Beta zeolite At longer
reaction times acrolein decomposed and acetaldehyde became the main product (max
selectivity ~26 mol in 30 min) The addition of metal-doped zeolites did not increase the
degree of glycerol conversion but increased the total selectivity towards the liquid products
48 wt La-doped NH4-Beta zeolite resulted in a 56 mol glycerol conversion with a 36 mol
selectivity towards acetaldehyde
The oxidation of glycerol with H2O2 in subcritical water was investigated in a
continuous fixed bed reactor at 125-175 degC 35 bar 60-300 s using H-Beta zeolite 25 wt-
and 48 wt Cu-doped H-Beta zeolite extruded with γ-Al2O3 As compared to non-catalysed
oxidation the addition of Cu-doped H-Beta zeolites did not increase the degree of conversion
but promoted the conversion rate of glycerol as well as the selectivity towards liquid products
The liquid products detected included dihydroxyacetone (DHA) formic acid (FA) acetic acid
glycolic acid pyruvaldehyde and lactic acid (LA) The distribution of these products varies
with the temperature residence time and the type of catalyst The top-three main products
vi
obtained were DHA FA and LA The highest yield of DHA (~8 mol) was achieved with 48
wt CuH-Betaγ-Al2O3 (MC) at 150 degC 60 s The same catalyst also provided LA with the
highest yield of 115 mol at 175 degC 240 s FA was detected with the highest yield of ~9
mol at 175 degC 60 s on H-Betaγ-Al2O3 (MC)
vii
Acknowledgments
First and foremost I sincerely thank my supervisor Prof Klaus Hellgardt for the
opportunity to work on a challenging project his guidance and encouragement throughout my
PhD studies I am also very grateful to Dr Radim Skapa for his advice feedback and support
provided throughout the last four years
I would like to thank Prof Kang Li and Dr Zhentao Wu for providing the ceramic
hollow fibres I am grateful to Dr Peter DiMaggio for allowing me to use the centrifuge
I am indebted to the people whose feedback and kind help have resulted in a substantial
contribution to my research project Dr Christos Kalamaras Dr John Brazier Dr Mimi Hii
Dr Pongsathorn Dechatiwongse Andrew Leung and Arash Izadpanah Special thanks for my
genius MSc students Cheok Neng Lucas Ho Mayuresh Patel and Lixin Song - Thanks guys
for all the smiles and laugh you put on my face
I have received great support from all of the members of the REaCT group Thank you
all for being very friendly and useful Dr Fessehaye Zemichael Dr Chun-Yee Cheng Dr
Oluseye Agbede Dr Franck Essiagne Dr Palang Bumroongsakulsawat Dr Maha Alsayegh
Dr Chin Kin Ong Dr Ju Zhu Bhavish Patel Muhammad Ibadurrohman Dr Lisa Kleiminger
Dr Sergio Lima Isaac Gentle Faye Al Hersh and Irina Harun
I really appreciate the tremendous support and fantastic caring warmth from Radim
Skapa thanks for all of his love inspiration and kindness I would also like to thank my friend
Wiparat Traisilanan for all her help all these years
Finally I sincerely and faithfully thank my family without whose love and support my
life would never been as wonderful as this
viii
Nomenclature
Symbol Meaning Usual Units
1H-NMR proton nuclear magnetic resonance spectroscopy
A pre-exponential factor s-1
BET Brunauer-Emmett-Teller
CHFM ceramic hollow fibre membranes
deAl-Beta dealuminated Beta zeolite
DHA dihydroxyacetone
Ɛ dielectric constant
Ɛ void fraction
Ea activation energy Jmol-1
EHS environment health and safety
FA formic acid
GC gas chromatography
GC-FID gas chromatography - flame ionisation detector
HLW hot liquid water
HMQC heteronuclear multiple quantum coherence
HPLC high-performance liquid chromatograph
HTW high-temperature water
ICP-OES inductively coupled plasma - optical emission
spectrometry
K equilibrium constant
k reaction rate constant s-1 (1st order reaction)
Kw ion dissociation constant of water
LA lactic acid
LCA life-cycle assessment
ix
MC methyl cellulose
MHz megahertz
mM millimolar
MPa megapascal
R universal gas constant (83145 JmiddotK-1middotmol-1)
RT room temperature
SCW supercritical water
SEM scanning electron microscopy
119878119864 standard error of mean
SSIE solid state ion exchange
sub-CW subcritical water
TEA triethanolamine
TEAOH tetraethylammonium hydroxide
TEM transmission electron microscopy
EDX energy-dispersive X-ray spectroscopy
TEOS tetraethylorthosilicate
TGA thermal gravimetric analysis
TPD temperature-programmed desorption
TUD-1 mesoporous silica matric named after the Technische
Universiteit Delft
UV-VIS ultraviolet-visible spectroscopy
XRD X-ray diffraction spectroscopy
XRF X-ray fluorescence spectroscopy
δ chemical shift ppm
ρ density kgmiddotm-3
All other abbreviations are described in the text as they occur
x
Table of Contents
Declaration iii
Abstract v
Acknowledgments vii
Nomenclature viii
Table of Contents x
1 Introduction 1
11 Glycerol conversion into value-added chemicals 3
12 Catalytic hydrothermal conversion as a process for converting glycerol 5
2 Literature review 9
21 Water under hydrothermal conditions 9
211 Applications of high temperature - high pressure water 9
212 Physicochemical properties of high temperature - high pressure water 10
213 Water and other green solvents 16
22 State of the art hydrothermal technology in the conversion of biomass and glycerol
helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip18
221 Production of liquid fuels by hydrothermal liquefaction of biomass and glycerol 20
2211 Biomass 20
2212 Glycerol 23
222 Production of gaseous fuels by hydrothermal gasification 23
2221 Biomass 25
2222 Glycerol 26
223 Production of bio-chemicals by hydrothermal conversion of biomass 33
224 Production of bio-chemicals by hydrothermal conversion of glycerol 42
2241 Non-catalytic hydrothermal conversion of glycerol 48
xi
2242 Hydrothermal conversion of glycerol catalysed by acids 51
2243 Hydrothermal conversion of glycerol catalysed by bases 56
2244 Hydrothermal conversion of glycerol with H2O2 62
23 Zeolite-based catalysts 65
231 Zeolites and their Broslashnsted and Lewis acidity 65
232 Stability of zeolites under hydrothermal conditions 67
233 Zeolite catalysts in the hydrothermal conversion of polyols to chemicals 74
234 Preparation of Lewis acidic zeolite catalysts 78
235 Immobilisation and shaping of powder zeolite catalysts 80
2351 Immobilisation of zeolites on inorganic membrane (zeolite membrane) 80
2352 Shaping of zeolite powder 83
2353 Agglomeration of zeolite powder 86
3 Aim of study 89
31 Aim 89
32 Objectives 90
4 Experimental protocols and methodology 94
41 Characterisation of solid samples 94
411 BrunauerndashEmmettndashTeller (BET) analysis 94
412 X-ray Fluorescence (XRF) spectroscopy 97
413 Powder X-ray diffraction (XRD) spectroscopy 99
414 Electron microscopy 101
415 Temperature programed desorption of ammonia (TPD-NH3) 103
416 Thermogravimetric analysis (TGA) and Differential scanning calorimetry (DSC) 106
42 Characterisation of liquid and gaseous samples 107
421 Nuclear magnetic resonance (NMR) spectroscopy 107
4211 Identification of liquid products 108
4212 Quantitative 1H-NMR analysis 112
xii
4213 Validation of KHP 115
422 High-performance liquid chromatography (HPLC) 116
4221 Quantitative analysis of results from HPLC vs 1H-NMR 118
423 Inductively Coupled Plasma - Optical Emission Spectrometry (ICP-OES) 118
424 Mass spectrometry (MS) 119
425 UV-VIS spectrophotometry 120
4251 Preparation of TiOSO4 solution 122
4252 Preparation of H2O2 stock solution 122
43 Chemicals and materials 124
44 Catalyst preparation 126
441 Preparation of H-Beta zeolite and H-ZSM-5 130
442 Preparation of metal-exchanged NH4-Beta H-Beta NH4-ZSM-5 and H-ZSM-5 zeolites
helliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphelliphellip131
443 Preparation of metal-exchanged dealuminated H-Beta zeolite 132
444 Preparation of CeH-Beta zeolite on ceramic hollow fibre membranes 133
445 Preparation of H-Beta and metal-dopedH-Beta zeolites on TUD-1 matrix 136
446 Preparation of extruded zeolite catalysts 137
45 Hydrothermal dehydration of glycerol 140
451 Dehydration of glycerol 140
452 Dissolution of H-Beta zeolite in hot liquid water 141
453 Effect of dissolved Al and Si on the dehydration of glycerol 142
46 Hydrothermal oxidation of glycerol 143
461 Batch process 143
462 Continuous flow process 144
5 Catalyst characterisation 155
51 H-Beta zeolite and H-ZSM-5 155
511 Powder X-ray diffraction 155
xiii
512 BET N2 analysis 160
513 TPD-NH3 163
52 Metal-exchanged zeolites 172
521 Powder X-ray diffraction 173
522 BET N2 analysis 178
523 TPD-NH3 180
53 Dealuminated H-ZSM-5 and H-Beta zeolite 184
531 Dealuminated ZSM-5 185
532 Dealuminated Beta zeolite 187
5321 Physical properties 187
5322 Surface acidity and the degree of Al extraction 194
533 Conclusions and discussion 196
54 Metal-exchanged dealuminated Beta zeolites 197
541 Physical properties 198
542 Surface acidity 209
55 Immobilisation and agglomeration of Beta zeolite-based catalysts 210
551 CeH-Beta zeolite on ceramic hollow fibre membranes 211
552 H-Beta and metal-dopedH-Beta zeolites embedded in TUD-1 matrix 216
553 Extruded catalysts 220
5531 Starting zeolite-based catalysts 220
5532 Beta zeolite-based catalysts extruded with bentonite clay 222
5533 Beta zeolite-based catalysts extruded with γ-Al2O3 225
6 Hydrothermal dehydration of glycerol 232
61 Non catalysed hydrothermal dehydration of glycerol 232
62 H-ZSM-5 and H-Beta zeolite catalysed hydrothermal dehydration of glycerol 238
621 Conversion degree and selectivity to liquid products 238
622 Composition of liquid products 246
xiv
623 Reaction kinetic analysis 249
63 Effect of lanthanide and transition metal doping 256
64 Hydrothermal stability of H-Beta zeolite 263
641 Dissolution of H-Beta zeolite 263
642 Effect of dissolved Si and Al on the dehydration of glycerol 269
643 Conclusion 280
7 Hydrothermal oxidation of glycerol with H2O2 281
71 Batch process 281
711 Non-catalysed oxidation 281
712 Zeolite-catalysed hydrothermal oxidation of glycerol 287
7121 H-Beta zeolite and H-ZSM-5 287
7122 Metal-exchanged Beta zeolite 291
72 Continuous flow process 297
721 Non-catalysed oxidation 298
73 Catalytic oxidation of glycerol 302
731 Effect of temperature on the catalytic oxidation of glycerol in subcritical water using H-
Betaγ-Al2O3 (MC) catalyst 302
732 Effect of residence time on the catalytic oxidation of glycerol in subcritical water using
H-Betaγ-Al2O3 (MC) catalyst 306
733 Effect of Cu-doped H-Betaγ-Al2O3 (MC) on the catalytic oxidation of glycerol in
subcritical water 307
7331 Conclusion 315
734 Stability of extruded Betaγ-Al2O3 (MC) catalysts 316
8 Summary and Future work 323
81 Summary 323
82 Future work 327
821 Hydrothermal dehydration of glycerol 328
xv
822 Hydrothermal oxidation of glycerol 328
9 References 329
10 Appendix 361
1
1 Introduction
In response to the ever-increasing price of fossil fuels as well as the environmental
issues associated with the use of these non-renewable resources the chemical industry has
focused its attention on alternative sustainable and renewable sources of fuels and chemicals
One example is the production of biodiesel by the transesterification of triglycerides the main
component of vegetable oils and animal fats with an alcohol see Figure 1
Figure 1 Biodiesel production process the transesterification of triglyceride (esters of
glycerol) with methanol yields fatty acid methyl esters and glycerol (adapted from Ciriminna
et al (2014))
Biodiesel is the second most used liquid biofuel after bioethanol The growth of the
biodiesel industry has been constant for many years (REN21 2014) Its global production was
about 26 billion litres in 2013 and it is projected to reach 40 billion litres in 2023 (OECDFAO
2014) A sharp increase of biodiesel production over the past decade resulted in a large surplus
of glycerol its main by-product as there is 1 kg of glycerol produced for every 9 kg of
biodiesel The production of glycerol by the biodiesel industry increased from 200000 tonnes
2
in 2004 (Mario Pagliaro and Rossi 2008) to more than 22 million tonnes in 2013 and it is
estimated to reach 35 million tonnes by 2023 (OECDFAO 2014) see Figure 2 The
overproduction of crude glycerol generated a problem of costly glycerol disposal making the
biodiesel production process economically less competitive On the other hand the
overproduction of glycerol resulted in a significant price drop (~ US $149tonne in November
2015 as compared to US $551tonne in 2004) (Pyle et al 2008 OPIS 2015) see Figure 2
making it an interesting feedstock for the production of value-added chemicals
00
05
10
15
20
25
30
35
Cru
de g
lycerol
[mil
lion
ton
ne]
Year
2004 2005 2006 2007 2008 2009 2010 2012 2014 2015 2023
0
100
200
300
400
500
Pric
e o
f cru
de g
lycerol (8
0
pu
re) [U
S $
ton
ne]
Figure 2 World crude glycerol production (bars) and its price (line) between 2004 and 2015
The crude glycerol production in 2004-2010 was reported by Almeida et al (2012) The
production in 2012-2023 was estimated based on the biodiesel production reported in
OECDFAO (2014) The crude glycerol (80 pure) prices were reported by Pyle et al (2008)
Clomburg and Gonzalez (2013) OPIS (2014) and OPIS (2015)
3
11 Glycerol conversion into value-added chemicals
The terms glycerol and glycerine (also called glycerin) are used interchangeably
although there are slight differences in their definitions and use (Bart et al 2010) Glycerol is
a pure chemical compound 123-propanetriol while glycerine is the commonly used
commercial name for products whose principal component is glycerol (typically ge 95) In
this work lsquoglycerolrsquo is used invariably without distinction between the two while the glycerine
obtained from the biodiesel process is specifically referred to as crude glycerol
Glycerol is non-toxic and highly hydrophilic due to the presence of three hydroxyl
groups which result in an extensive network of H-bonding (Katryniok et al 2009) It is one
of the top twelve platform chemicals identified by the United States Department of Energy
(DoE) (Werpy and Petersen 2004) and its potential to become a primary raw material for bio-
refineries has been demonstrated by numerous studies almost 5000 scientific papers
concerning the conversion of glycerol into value-added chemicals were published in peer-
reviewed scientific journals between 2000 and 2015 (Googlescholarcouk 2015)
In addition to its traditional uses in food pharmaceutical and cosmetic industries the
highly functionalised molecule of glycerol may be converted into a variety of higher value
compounds by numerous different conversion processes see Figure 3 few of which have been
commercialised (Katryniok et al 2011)
chlorination for the production of epichlorohydrin
gasification to yield syngas used for the manufacture of bio-methanol see Chapter
2222
indirect esterification for the manufacture of mono- and diacylglycerols and
fermentation for the production of dihydroxyacetone (DHA) see Chapter 224
4
A number of processes have not been transposed to the industrial scale due to their poor
economies associated with low selectivity andor high operation costs (Kongjao et al 2011)
One example is the catalytic conversion of glycerol under hydrothermal conditions The
hydrothermal conversion process gained significant attention in the last two decades (Peterson
et al 2008) due to its potential to overcome the difficulties associated with other methods of
glycerol transformations such as the long reaction time of fermentation
Figure 3 The known conversion processes of glycerol into value-added chemicals (adapted
from Pagliaro et al (2007) Zhou et al (2008) and Katryniok et al (2011))
Finding a commercially viable application of the waste glycerol would not only make
biodiesel production economically more competitive but also decrease the dependence of
5
society on fossil fuels This work focuses on catalytic hydrothermal conversion as a process
for converting glycerol into value-added chemicals
12 Catalytic hydrothermal conversion as a process for
converting glycerol
Hydrothermal technology has been investigated since the 1970rsquos (Bonn et al 1983)
and widely used for pre-treatment of biomass materials containing water such as
lignocellulosic biomass agricultural residues and municipal wastes (Peterson et al 2008) It
is considered to be one of the most suitable methods for processing of crude glycerol for the
following reasons (Peterson et al 2008)
The water present in crude glycerol see Table 1 for composition of crude glycerol
acts as a medium in hydrothermolysis eliminating expensive drying required for
other conversion methods
Hydrothermal technology is capable of processing mixed feedstocks containing
both organic and inorganic compounds eliminating the need for costly purification
There is no need to maintain microorganisms as required for fermentation process
The process efficiency may be controlled by adjusting the temperature and pressure
6
Table 1 The composition of crude glycerol produced as a by-product of biodiesel industry
(adapted from Ye and Ren (2014))
Composition [wt] Crude glycerol
Sample A1 Sample B2 Sample C3
glycerol 65-85 229-333 ~83
ash 4-6 27-30 Not determined
methanol 234-375 86-126 lt1
water 1-3 41-182 ~13
sodium 01-4 16-19 lt5
soap Not determined 205-262 Not determined
fatty acids methyl esters Not determined 193-288 Not determined
glycerides Not determined 12-7 Not determined
free fatty acids Not determined 10-30 Not determined 1 Crude glycerol produced from batch process in lab (Thompson and He 2006 Ciriminna et
al 2014) 2 Crude glycerol from PolyGreen Technologies LLC (Mansfield OH) (Hu et al
2012) 3 Crude glycerol from Cargillrsquos biodiesel plant (Kansas City MO)
It is known that water under hydrothermal conditions acts as a highly reactive acid-base
catalyst allowing short reaction times (Akiya and Savage 2002) The selectivity to desired
conversion products under hydrothermal conditions may be controlled by
(i) adjusting the temperature and pressure influencing the properties of subcritical
water (sub-CW) and supercritical water (SCW) andor
(ii) adding an additive or a catalyst
By controlling the nature of additives andor catalysts the glycerol typically undergoes
one of the following conversion reactions under hydrothermal conditions
acid-catalysed dehydration to acrolein
base-catalysed dehydrogenation followed by dehydrationrehydration to lactic acid and
7
wet oxidation using an oxidant to form low molecular weight organic acids
Inorganic acids and bases were reported to exhibit high performance The hydrothermal
dehydration of glycerol at 300-400 ordmC in the presence of H2SO4 or NaHSO4 resulted in ~22-86
mol yield of acrolein (Antal et al 1985 Ramayya et al 1987 Watanabe et al 2007) while
the conversion of glycerol in subcritical water (280-300 ordmC) using alkaline hydroxides such as
NaOH and KOH as catalysts led to high yields (85-90 mol) of lactate (metal salt of lactic
acid) (Kishida et al 2005 Shen et al 2009 Ramirez-Lopez et al 2010) However the use
of strong inorganic acids or bases has raised concerns over their corrosive effect on the
equipment prompting the research on alternative catalysts
The addition of noble or transition metal catalysts ie Ru Pt Au Rh Ir and Cu was
found to effectively catalyse the dehydrogenation of glycerol under alkaline hydrothermal
conditions resulting in comparative lactate yields at temperature as low as 200 degC as compared
to 300 degC required for alkaline hydrothermal conversion (Maris and Davis 2007 Roy et al
2011 Auneau et al 2012) This process is however not without disadvantages since the
resulting lactate needed to be converted into free lactic acid by neutralizing with a strong acid
eg H2SO4 generating an issue associated with the separation and disposal of metal sulfates
Solid acid catalysts have recently gained significant interest as an alternative to mineral
acids for the dehydration of glycerol Solids acid catalysts are solids that possess acidic
functional groups on their surfaces functioning as catalysts just like water soluble acids such
as H2SO4 or HCl (Hattori and Ono 2015) Solid acid catalysts such as RuZrO2 and WO3TiO2
were demonstrated to have promising activity providing acrolein yield of approximately 50
mol (May et al 2010 Akizuki and Oshima 2012)
Zeolites are solid catalysts exhibiting both Broslashnsted and Lewis acidity promoting the
dehydration and dehydrogenation of glycerol respectively The high performance of zeolites
8
and metal-doped zeolites also known as Lewis-acidic zeolites in the conversion of glycerol
and other polyols to value-added chemicals has been successfully demonstrated by several
research groups (Taarning et al 2009 de Oliveira et al 2011 Gonzalez-Rivera et al 2014)
In spite of the fact that zeolites generally have limited stability under hydrothermal conditions
many studies reported the use of zeolites as catalysts under such conditions However more
work needs to be carried out to improve the understanding of hydrothermal conversion of
glycerol in the presence of zeolites
9
2 Literature review
This chapter presents a comprehensive review on the properties of water under
hydrothermal conditions followed by its applications in biomass conversion with special
reference to the dehydration dehydrogenation and oxidation of glycerol in aqueous phase
(Chapter 224) Finally the potential of zeolites as catalysts in the hydrothermal conversion of
glycerol and their modification are discussed in Chapter 23
21 Water under hydrothermal conditions
211 Applications of high temperature - high pressure water
Water often referred to as ldquogreen solventrdquo is one of the most environmentally and
economically attractive solvents because it is renewable non-toxic inflammable readily
accessible and inexpensive However water is a poor solvent for numerous organic compounds
reactions reducing its potential for replacing conventional non-aqueous non-polar solvents
derived from petroleum crude oils On the other hand at high temperature - high pressure
water has the properties enabling it to dissolve non-polar compounds
A thermochemical process in which water at high temperature - high pressure (HTP-
water) is used as a reaction medium is called hydrothermal process The hydrothermal medium
is also referred to as subcritical water (sub-CW) amp supercritical water (SCW) hot-compressed
water (HCW) pressurised-hot water (PHW) or superheated water in the present work
The HTP-water may not only be used as a solvent with a broad range of applications
but also as a catalyst in chemical synthesis biomass conversion coal liquefaction plastics
10
recycling and waste destruction (Calvo and Vallejo 2002 Galkin and Lunin 2005 Peterson
et al 2008 Hayashi and Hakuta 2010) It may also be used as a medium for instance to
extractisolate natural products (Herrero et al 2006) heavy metals hydrocarbons or polycyclic
aromatic hydrocarbons (PAHs) from contaminated soils (Hawthorne et al 1994 Kronholm et
al 2002) The use of HTP-water in the above mentioned processes is mainly driven by the
benefit of more environmentally friendly economical and safer processes
212 Physicochemical properties of high temperature - high
pressure water
The physicochemical properties of water vary significantly with pressure and
temperature see Figure 4 for the pressure - temperature (PT) diagram of water The critical
point of water lies at 374 degC and 2206 bar marking the transition point between the subcritical
and supercritical regions While supercritical water (SCW) is typically defined as water above
its critical temperature and pressure the temperatures and pressures used to define subcritical
water (sub-CW) vary widely in different studies For instance Ramos et al (2002) defines the
sub-CW by the temperatures between 100 and 374 degC and pressures high enough to maintain
its liquid state Yu et al (2008) defines the sub-CW region by the temperatures above 150 degC
and various pressures whereas water at the temperature ranging from 100 to 360 degC at saturated
vapour pressure is a definition adopted by Pourali et al (2010) Many works on hydrothermal
biomass processing arbitrarily place the sub-CW region between 200 degC and 374 degC (Mok and
Antal 1992 Minowa et al 1998 Peterson et al 2008) In this work the sub-CW region is
defined by the temperatures ranging between 100 degC and 374 degC (Clifford 2007 Cheng and
Ye 2010) and the pressure equals to or is higher than saturated vapour pressure of water
11
Figure 4 The pressure - temperature (PT) diagram of water with highlighted subcritical and
supercritical regions
As the temperature increases the number of hydrogen bonds between the water
molecules decreases reducing the dielectric constant and thus lowering the polarity (Mountain
1989 Mizan et al 1996) As a result the solubility of organic compounds in HTP-water is
much higher than in water under ambient conditions The selected physicochemical properties
of water such as density (ρ) dielectric constant (Ɛ) and ion dissociation constant (Kw) are
shown in Figure 5 as a function of temperature The values of all three physical properties
decrease with increasing temperature The density continuously decreases from 1000 kgmiddotm-3 at
25 degC to 600 kgmiddotm-3 at 350 degC and then further drops to 150 kgmiddotm-3 at 450 degC The dielectric
constant falls from 80 to less than 2 as the temperature rises from 25 degC to 450 degC The ion
12
dissociation constant initially increases from 10-14 to 10-11 at just below 350 degC and then
plummets to 10-175 above 500 degC
Figure 5 Static dielectric constant (Ɛ) density (ρ) and ion dissociation constant (Kw) of water
at 300 bars as a function of temperature The values of all three physical properties decrease
with increasing temperature The figure was taken from Peterson et al (2008)
Several physicochemical properties of HTP-water are similar to those of common
organic solvents at ambient conditions (Bialkowski Smallwood 1996) Table 2 summarises
the characteristics of water under different conditions compared to acetone
13
Table 2 Characteristics of water under different conditions compared to acetone a common
organic solvent at ambient conditions (adapted from (Krammer and Vogel 2000) Galkin and
Lunin (2005) and Moumlllerthinsp et al (2011)) L liquid phase G gas phase
Characteristic Ambient
Water
Subcritical
Water
Supercritical
Water
Acetone
Temperature [degC] 0-100 100-374 gt374 RT
Vapour pressure
[bar]
003 (24 degC) 1 (100 degC) to 221
(374 degC)
gt221 024
(20 degC)
Aggregate state liquid liquid no phase separation liquid
Density [kg m-3] 997 (25 degC) 958 (101 degC 11
bar) 692 (330 degC
300 bar)
between gas-like and
liquid-like densities eg
252 at140 degC 300 bar
7899 (20
degC 11
bar)
Viscosity [microPas] L 884 G 99
(25 degC)
L 277 G123
(101 degC) L504
G 307 (371 degC)
low 330
(25 degC)
Heat capacity Cp
[kJ kg-1K-1]
422 (25 degC) 486 13 (400 degC 250 bar) 218
(20 degC)
Dielectric
constant
785 (25 degC 1
bar)
271 (250 degC 50
bar) 182 (330 degC
300 bar)
59 (400 degC 250 bar)
105 (400 degC 500 bar)
206
(20 degC)
Compressibility no slightly increased
but still a liquid
(at 370 degC)
yes na
-logKw 14 na 21 na
(Bialkowski Smallwood 1996) (Kerton 2009) Note L liquid phase G gas phase
The ionic product of water Kw determining the concentration of hydronium (H3O+)
and hydroxide (OH-) ions (see Figure 6) is another important factor describing the properties
of water The Kw of sub-CW at 350 degC is nearly 1000 times higher than that of ambient water
see Figure 5 Thus the higher concentration of H3O+ leads to much lower pH see Figure 7
meaning that the point of neutrality shifts from ~7 to ~55 In contrast the concentrations of
H3O+ and OH- in SCW are much lower than in ambient water resulting in pH of above 7 but
14
the free-radical chemistry starts to play more important role in this region (Akiya and Savage
2002) The properties of HTP-water are reviewed extensively by Akiya and Savage (2002)
Figure 6 The relationship between the water ion product (Kw) and the pH
Figure 7 pH of water as a function of temperature and pressure (taken form Cook and Olive
(2012))
15
The hydrogen bonding influences several important characteristics of water including
the dielectric constant polarity viscosity and diffusivity These properties change with density
which is greatly dependent upon temperature and pressure Figure 8 shows the temperature and
pressure dependence of the density of water
Figure 8 The temperature and pressure dependence of the density of water under hydrothermal
conditions (taken from Cook and Olive (2012))
Water in near-critical and supercritical region exhibits the properties of nearly an ideal
gas and thus the equation of state of an ideal gas Pv = nRT may also be used for near-critical
and SCW (Brunner 2009) The vapour pressure (P) volume (V) temperature (T) and the
transport properties eg diffusivity and viscosity of near-critical and SCW are similar to that
16
of supercritical fluids1 but the density of the former is slightly higher and is very sensitive to
minor changes in temperature and pressure
The solubility of most gases in water tends to decrease with increasing temperature
(Akiya and Savage 2002) however as the solubility reaches its minimum eg around 100 degC
for O2 (Battino et al 1983) it starts to increase at increased temperature and pressure The
solubility of gases and non-polar compounds in sub-CW increases with increasing temperature
reaching complete miscibility with SCW (Brunner 2009) SCW acts as a non-polar solvent
such as hexane (Mizan et al 1996 Akiya and Savage 2002) On the other hand the solubility
of inorganic salts in near-critical and SCW is much decreased (Mizan et al 1996)
213 Water and other green solvents
A large amount of organic solvents is used each year in the chemical industry for their
good heat amp mass transfer and diffusion of reactants However the environmental and health
issues associated with organic solvents have prompted the research into the use of more
environmentally-friendly (green) solvents
The environmental footprint of solvents is most commonly assessed by (i) the
environment health and safety (EHS) method and (ii) the life-cycle assessment (LCA) method
(Capello et al 2007 Jessop 2011) The EHS method identifies the impact of a solvent on the
environment while the LCA method assesses the emissions to the environment as well as the
full life-cycle of a solvent
1 A supercritical fluid is any substance at above its critical temperature and pressure ie a temperature and pressure
above which distinct liquid and gas phases do not exist (Carlegraves 2010)
17
Water is the most environmentally friendly solvent with no negative impact on the
environment human health and safety It is also believed according to a survey by Jessop
(2011) to be the second best option after supercritical CO2 for reducing solvent-related
environmental damage Zhang et al (2008) compared an ionic liquid ([bmim]BF4) water
acetone benzene and lithium perchlorate diethyl ethers mixture (LPDE) in terms of
environmental impact by 11 factors including the impact of both the solvent itself and its
synthesis It was found that water has the lowest while the ionic liquid the highest
environmental impact in every category except ozone-layer depletion in which benzene
dominated Water is also the greenest solvent among those listed in the GlaxoSmithKline
(GSK) solvent guide for medicinal chemistry see Table 3 The guide highlights potential issues
of each solvent and provides possible alternatives It also includes life cycle assessment
information
Table 3 Selected examples of solvents listed in the GlaxoSmithKline (GSK) solvent guide for
medicinal chemistry (Henderson et al 2011)
Solvent MP
[ordmC]
BP
[ordmC]
Waste Environmental
impact
Health Flammability
amp Explosion
Reactivity
Stability
Life cycle
score
Legislation
Flag
EHS
Red Flag
Water 0 100 4 10 10 10 10 10
2-Propanol -88 82 3 9 8 6 8 4
1-Butanol -89 118 5 7 5 8 9 5
Ethanol -114 78 3 8 8 6 9 9
Methanol -98 65 4 9 5 5 10 9
Acetone -95 56 3 9 8 4 9 7
Acetic acid 17 118 4 8 6 8 7 8
Toluene -95 111 6 3 4 4 10 7
Heptane -91 98 6 3 8 3 10 7
xxxxxxxxxx Substitution recommended ndash There are no current restrictions but future regulatory restrictions may apply
xxxxxxxxxx Must be substituted ndash A regulatory ban applies
Note MP = melting point BP = boiling point EHS = environment health and safety
18
Despite several advantages the use of water is not without drawbacks as reflected by
the low score (only 4) in the waste recyclingtreatment in Table 3 also see Table 4 Firstly
water requires energy intensive distillation Secondly the production of a contaminated
aqueous waste stream may have significant environmental and economic impacts (Lancaster
2002) for example the concentration of contaminated water by distillation is energy intensive
as compared to concentration of a propanol waste stream Finally high specific heat capacity
of water makes it difficult to heat or cool rapidly
Table 4 Advantages and disadvantages of water as a solvent (adapted from Lancaster (2002))
Advantages Disadvantages
Non-toxic Distillation is energy intensive
Opportunity for replacing volatile organic
solvents
Contaminated waste streams may be difficult to
treat
Naturally occurring High specific heat capacity - difficult to heat or
cool rapidly
Inexpensive
Non-flammable
High specific heat capacity - exothermic
reactions can be more safely controlled
22 State of the art hydrothermal technology in the
conversion of biomass and glycerol
Hydrothermal processing is considered to be one of the most suitable technologies for
treatment of wet biomass and biomass-derived glycerol due to the intrinsic water content of
such feedstocks and the unique solvent properties of SCW (see Chapter 21) The technology
19
has been long used in processing of non-food plant biomass such as agricultural and municipal
wastes (Peterson et al 2008 Brunner 2009) and is expected to play an important role in
future production of fuels and chemicals from biomass This chapter reviews the state of the
art of the technology as a process for producing biofuels and biochemical utilizing both general
biomass materials as well as bio-derived glycerol as source materials
Hydrothermal processing of biomass may not only yield gaseous- but also liquid- and
solid biofuels Depending on the temperature and pressure the biomass conversion may
progress via see Figure 9 (Peterson et al 2008)
liquefaction (typically 200-370 degC 40-200 bar)
catalytic gasification (near-critical temperatures up to 500 degC) and
high-temperature gasification (typically ge500 degC)
The processes above are discussed in Chapters 221 and 222
20
Figure 9 Hydrothermal processing regions referenced to the pressure ndash temperature phase
diagram of water The figure was taken from Peterson et al (2008)
221 Production of liquid fuels by hydrothermal liquefaction of
biomass and glycerol
2211 Biomass
Hydrothermal liquefaction also known as hydrous pyrolysis (Obeid et al 2015) is a
thermochemical conversion process in which a high temperature (typically 200-370 degC) - high
pressure (50-400 bar) water is used to break the chemical bonds of biomass material producing
21
liquid biofuels such as bio-oil (Peterson et al 2008 Zhang 2010) In contrast to pyrolysis2 it
involves direct liquefaction of wet biomass into bio-crude (also known as liquid bio-oil) which
may be further processed to yield transport fuels andor various chemicals The potential of
biomass as a source of renewable fuels was realized as early as in the first half of the 20th
century when Berl (1944) reported that a petroleum-like product could be obtained by alkaline
hydrothermolysis (~230 degC) of feedstocks such as cornstalks sugarcane seaweed algae and
grass
Biomass materials typically contain 30-50 wt oxygen and have a heating value of 10
to 20 MJkg-1 (GCEP 2005) A liquefaction process involves contacting the biomass with high
temperaturepressure water (typically 200-370 degC 40-250 bar) to yield a bio-crude an oily
liquid with reduced oxygen content (10-20 wt) and a significantly increased heating value of
30 to 37 MJ kg-1 (Toor et al 2011) However the oxygen content of the bio-crude is still much
higher than that of petro-derived oil (lt1 ) (Aitani 2004) The high oxygen content of the
biomass liquefaction product results in undesirable properties such as lower energy capacity
lower volatility and reduced thermal stability Therefore it is desirable to further reduce the
oxygen content by its conversion either into CO2 or more desirably H2O
An increased interest of researchers in the liquefaction of cellulosic biomass was
recorded during the first oil crisis in 1970s A considerable amount of pioneering work was
done by a group of researchers at the Pittsburgh Energy Research Centre (PERC) of the United
States Bureau of Mines Appell et al (1971) developed a continuous flow process for
liquefaction of wood chips into heavy oil using recycled oil as a reaction medium and a source
of hydrogen at 330-370 degC and 200 bar in the presence of CO as a reducing gas and Na2CO3
as a catalyst Although a moderate oil yield was obtained (45-55 relative to the mass of dry
2 Pyrolysis is an anaerobic thermochemical decomposition process of dried biomass yielding solid gaseous and
liquid fuels (Demirbas 2001)
22
biomass used) the need for a large amount of recycled oil as a source of hydrogen is not
practical The use of recycled oil is not required in the process developed by the researchers at
Lawrence Berkeley Laboratory whose liquefaction process utilizes hydrothermal media
allowing the reactions to take place in a water-rich phase at high pressure (Peterson et al
2008) In this process the biomass is hydrolysed at 330 to 360 degC 100 to 240 bar in the presence
of H2SO4 followed by neutralization with Na2CO3 These two processes were demonstrated on
a pilot plant scale at Albany Oregon facility but further research was halted due to financial
reasons as well as the lack of basic scientific understanding of the process (Behrendt et al
2008) In 1982 the researchers at the Royal Dutch Shell Laboratory started to develop a so
called hydrothermal upgrading (HTU) process (Okkerse and van Bekkum 1999 Toor et al
2011) ie an biomass feedstock was treated in sub-CW at 300-350 degC 100-180 bar yielding a
bio-crude (Goudriaan et al 2005) The crude oil readily separates from water and due to its
low oxygen content it may be further upgraded cost-effectively by hydrodeoxygenation to a
clean diesel-type fuel with high cetane number (Okkerse and van Bekkum 1999) Additionally
in contrast to fossil fuels bio-crude does not require removal of inorganic contaminants such
as nitrogen and sulphur The research was abandoned in 1988 by the oil company but resumed
in 1997 by a Dutch consortium
Since the publication of the pioneering work on liquefaction of biomass by Appell et
al (1971) at the Pittsburgh Energy Research Centre (PERC) a number of research studies on
hydrothermal conversion of lignocellulosic biomass into liquid (bio-crude) gaseous and solid
fuels have been published up until July 2015 about 6700 articles concerning the hydrothermal
liquefaction were published in peer-reviewed online journals (Googlescholarcouk 2015)
However despite the intensive investigation so far all the existing works have been carried out
on a laboratory or bench-scale with only a few on demonstration plants (Toor et al 2011) The
hydrothermal technologies for liquefaction of biomass were reviewed by Behrendt et al
23
(2008) Peterson et al (2008) and Toor et al (2011) The proposed mechanisms of liquefaction
were reviewed by Demirbas (2000)
2212 Glycerol
The hydrothermal treatment of glycerol in the liquefaction region ie 200-370 degC and
40-200 bar does not result in an oil phase but rather in water-soluble compounds such as
acrolein acetaldehyde and acetic acid see Chapter 223 To the best of our knowledge a single
study reported a waxy-oil phase from hydrothermal conversion of glycerol Onwudili and
Williams (2010) attained ~62 wt yield of waxy-oil by non-catalytic hydrothermal
gasification (300-450 degC saturated vapour pressure) of crude glycerol containing 21 wt
methanol 42 wt glycerol and 33 wt fatty acid methyl esters As the temperature increased
from 300 degC to SCW conditions the waxy product transformed into a liquid oil at ge 380 degC
The formation of the waxyoil products may possibly be due to the presence of fatty acid methyl
esters in the crude glycerol as speculated by Leow et al (2015) who investigated the
hydrothermal liquefaction (HTL) of microalgae and found that the bio-crude yield and the
carbon distribution increased in proportion to the fatty acid content in microalgae feedstock
222 Production of gaseous fuels by hydrothermal gasification
Gasification is commonly defined as an aerobic thermochemical conversion process in
which dried biomass materials (lt 10 wt moisture) contacted with a controlled amount of
oxygen andor steam at high temperature (ge 371 degC) yield mainly a mixture of CO and H2 (also
known as synthesis gas or syngas) (Demirbas 2001 Elliott 2008) The requirement for dried
biomass results in high operating cost since the drying process is an energy- and time-
24
consuming step This may be avoided by gasification of wet biomass in sub-CW amp SCW see
Equation 1
Equation 1 Hydrothermal gasification of cellulose (C6H10O5) in SCW
Hydrothernal gasification may be applied in two practically possible scenarios known
as
(i) catalytic gasification and
(ii) high-temperature gasification
The former usually takes place in sub-CW or SCW at up to 500 degC with active metal catalyst
while the latter requires the temperatures above 500 degC with or without a catalyst typically
the activated carbon or alkali salts are used to inhibit the tarchar formation (Kruse et al 2000
Schmieder et al 2000 Osada et al 2006) Each method has both advantages and
disadvantages The catalytic gasification in SCW (374-500 degC) requires a transition metal
catalysts such as nickel and ruthenium whilst a highly active metal catalyst such as Raney
nickel or platinum with low weight hourly space velocities are usually used in sub-CW (Elliott
et al 1993 Davda and Dumesic 2004 Elliott et al 2006) Complete gasification of
lignocellulosic materials was reported using the catalytic gasification however the catalyst
deactivation is an issue and needs further optimisation On the other hand high-temperature
gasification usually leads to high feed conversion with high CO concentration due to low rate
25
of water-gas shift reaction but is energy intensive (Azadi and Farnood 2011) The SCW
gasification is reviewed extensively by Kruse (2008)
2221 Biomass
Hydrothermal gasification has been an intensive research topic for the last two decades
Its potential as an effective process to convert biomass to gaseous fuels was not realised until
Modell (1985) reported that maple wood sawdust treated in SCW (374 degC 220 bar) rapidly
decomposed to tar and gas composed mainly of H2 CO2 and CO without the formation of char
which represent one of the main issues in steam reforming usually carried out at ge 600 degC under
atmospheric pressure (Matsumura et al 2005) A study by Sealock and Elliott (1991) showed
that the degree of hydrothermal gasification of a wide range of lignocellulosic materials at 300
degC and ge101 bar was substantially promoted by the presence of a Ni metal or alkali-promoted
Ni catalyst ~40 gas yield comprising of CO2 H2 and CH4 was obtained The effect of alkali
on the carbon conversion to gases decreased in the following order Cs gt K gt Na Elliott et al
(1993) systematically evaluated the catalytic performance of various base metals3 and their
combinations noble metals and numerous support compositions for the process The metals
tested for gasification of biomass at 350 degC and autogenous pressure from 170-230 bar included
Co Cr Cu Mo Ni Pd Pt Rh Ru and Zn however only Ni Ru and Rh were found to be
active The supports stable in hot-compressed water were α-alumina and zirconia whereas other
types of alumina tested reacted with water to form boehmite losing the surface area and
physical strength Silica and titania were not susceptible to the hot-compressed water but their
tableted forms disintegrated
3The term ldquobase metalsrdquo (Smartmcx) refers to the metals that are not considered precious such as copper lead
nickel tin or zinc and are usually used in commercial and industrial applications for their high abundance and
much lower price as compared to precious metals
26
Minowa et al (1995) investigated the gasification of wet cellulose in sub-CW (350 degC
180 bar) using a reduced Ni catalyst The yield of gaseous products (H2 CO2 and CH4)
increased with the catalyst loading reaching 91 wt when 20 wt catalyst was used The high
gas yields obtained were in accordance with expectations since the Ni catalysts are well-known
highly performing catalysts however the nickel oxide was not found to be an active catalyst for
the SCW gasification of biomass (Modell 1985 Yamamura et al 2009) Other metal oxides
investigated included MoO3 ZrO2 and RuO2 all of which only RuO2 showed promising
activity and only in high temperature SCW (Izumizaki et al 2005 Yamamura et al 2009) In
addition to the reduced Ni and Raney (skeleton) Ni catalysts other reported active metal
catalysts for gasification in low-temperature subcritical water also known as aqueous phase
reforming (APR) were Pt-black PtTiO2 PtC and PtAl2O3 (Shabaker et al 2003)
Modification of commercially available industrial catalysts was reported to increase
their hydrothermal stability and activity in hydrothermal gasification Elliott et al (2006)
incorporated Ru Cu Ag Re or Sn into a BASF reforming catalyst increasing its stability in
hydrothermal environment Ru-incorporated BASF catalyst exhibited the highest stability and
activity for gasification and methanation Similarly the surface modification of Raney Ni with
small quantities of Sn was found to significantly enhance the H2 to CH4 ratio of the products
(Huber et al 2003 Shabaker and Dumesic 2004 Shabaker et al 2004 Shabaker et al 2005)
2222 Glycerol
Glycerol is an ideal feedstock for SCW gasification as it yields H2 and CO2 without the
coke formation see Equation 2 (Antal et al 2000 Buumlhler et al 2002 van Rossum et al
2009) Most studies on SCW gasification of glycerol aim to produce H2 a clean energy carrier
27
Equation 2 The SCW gasification of glycerol
Antal et al (1985) was first to study a non-catalysed glycerol decomposition in SCW
at 500 degC and 345 bar obtaining acetaldehyde acrolein and a gas mixture consisting of H2
CO2 CO CH4 C2H4 and C2H6 About 32 mol conversion of glycerol was obtained in ~15
min with H2 and CO2 as the main gaseous products possibly formed by the so called water-
gas shift reaction see Equation 3
Equation 3 Water-gas shift reaction
119862119874 + 1198672119874 harr 1198621198742 + 1198672
The non-catalytic SCW gasification of glycerol was also studied by Kersten et al
(2006) who reported a significant effect of temperature on the degree of conversion up to 650
degC Above this temperature the degree of conversion was a function of feed concentration and
only diluted glycerol solution was successfully completely converted to gas phase
Byrd et al (2008) demonstrated that near-theoretical H2 yield ie 7 mol H2mol
glycerol see Equation 2 was achieved by reforming 5 wt glycerol solution in SCW over a
RuAl2O3 catalyst at 800 degC 241 bar Although a high H2 yield was obtained the process
requires high energy input On the other hand May et al (2010) found that the SCW
28
gasification of 5 wt glycerol solution at 550 degC 350 bar led to a very low H2 yield less than
06 mol H2mol glycerol in the presence of RuZrO2
Chakinala et al (2010) hydrothermally gassified glycerol forming the backbone of
lipids (present up to 40 wt) in algae at 550-650 degC 250 bar 5 s residence time using a
continuous flow tubular reactor The effect of (i) amino acid (glycine L-alanine and L-proline)
and (ii) alkali salt (K2CO3) on the degree of glycerol conversion and the composition of its
conversion products were also tested The presence of amino acid was found to induce coke
formation whilst the addition of K2CO3 enhanced the gasification and promoted the H2
formation (~261 mol H2mol glycerol vs ~184 mol H2mol glycerol in a non-catalytic reaction
under identical conditions 600 degC 250 bar 5 s) The addition of alkaline catalysts is known to
inhibit char formation in SCW conversion of biomass resulting in higher oil andor gas yields
(Minowa et al 1998 Kruse et al 2000)
The majority of works on SCW gasification of glycerol were carried out on pure diluted
glycerol solutions Onwudili and Williams (2010) studied the hydrothermal gasification of
crude glycerol containing 21 wt methanol 42 wt glycerol and 33 wt fatty acid methyl
esters The non-catalysed reforming reactions carried out in a Hastelloy-C batch reactor at 300-
450 degC autogeneous pressure of 85-310 bar resulted in the products containing ~62 wt
oilwax which became liquid oil at ge380 degC ~10-19 wt of water-soluble products and ~3-18
wt of gases composed of H2 CO CO2 and light hydrocarbons were also produced When
NaOH was used as a catalyst at 380 degC a gas containing of up to 90 vol H2 ie an equivalent
to ~19 mol of H2 per mole of carbon in the feed was obtained The obtained H2 yield was
higher than that from pure glycerol under identical conditions No char formation was
observed The effect of high feed concentration (10-50 wt) and various alkaline catalysts
were also investigated at 445-600 degC 250 bar a residence time of 39-90 s by Guo et al
(2012) The gas yield increased with temperature and the gasification efficiency decreased with
29
increasing glycerol concentration The addition of alkali catalysts at the weight ratio of glycerol
to catalyst ~20 greatly enhanced the water-gas shift reaction increasing the H2 yield in the
following order K2CO3lt KOHlt Na2CO3ltNaOH The addition of a small amount of NaOH
(01 wt) led to the highest H2 yield of 493 molmol glycerol at 526 degC No char or tar
formation was observed in the experiments
The gasification in sub-CW also known as aqueous-phase reforming (APR) developed
by Cortright et al (2002) is another effective process for producing H2 from sugars and
alcohols including glycerol At 265 degC 56 bar 99 conversion of glycerol (1 wt solution)
with ~51 of selectivity to H2 was achieved over a Ptγ-Al2O3 catalyst (Cortright et al 2002)
A complete conversion of glycerol with 76 selectivity to H2 at 265 degC 514 bar was achieved
using inexpensive Raney Ni-Sn catalyst (Huber et al 2003) In addition to a lower
temperature as compared to typical hydrothermal gasification the APR yields less CO due to
the water-gas shift reaction see Equation 3 The works including the experimental conditions
and results reported on APR of glycerol are summarised in Table 5
30
Table 5 The list of works including the experimental conditions and results reported on aqueous phase reforming (APR) of glycerol The
percentages reported are in mol unless otherwise stated Note X = glycerol conversion S = selectivity Y = yield
[glycerol] Catalyst Reaction conditions
Results Ref Temp [degC] Press [bar]
[1 wt] Pt γ- Al2O3 265 56 X = 99 S (H2) = 51 Cortright et al (2002)
[1 wt] Raney Ni-Sn 265 514 X = 100 S (H2) = 76 Huber et al (2003)
[10 wt pure amp crude] 3 wt Ptvarious
Al2O3 - with different Pt precursors
250 20 (argon) Pt on a mixed γ- θ- δ- Al2O3 X = 45
S (H2) = 85
Lehnert and Claus (2008)
[5 wt amp10 wt] 03-12 wt Ptγ-Al2O3 180 - 220 114 - 25 X = 65 Y (H2) = 45 (at 5 wt glycerol) Luo et al (2008)
[1 wt] Niγ-Al2O3 (with Ce La Mg Zr
promoters)
225 30 La-Niγ-Al2O3 was most active X = 37 Y not
reported gas phase composition 32 H2 40
CO2 28 CH4
Iriondo et al (2008)
[10 wt] Pt Ni Co Cu on Al2O3 SAPO-
11 MgO H-USY SiO2 active C
230 32 PtAl2O3 was most active X = 19 Y not reported
70 H2 present in gas phase
Wen et al (2008)
[5 wt] Ni-Coγ- Al2O3 220 25 X = 45 Y (H2) = 40 Ce doping enhanced
catalytic activity
Luo et al (2010)
[10 wt] 1-3wt Re-3wt PtC 225 29 3wtRendash3wtRtC was most active X = ~88 S
(gas) = ~55 KOH increased H2 productivity and
selectivity to oxygenates in aqueous phase
King et al (2010)
[1 wt] Pt on Al2O3 ZrO2 MgO CeO2 225 autogenous X = 13-26 Y not reported 62-72 H2 in gas
phase
Menezes et al (2011)
[1 wt amp 10wt]
NiCeO2
250 270 autogenous
(375 527)
X = ~90 Y not reported ~90 H2 present in gas
phase the rest included CO CO2 and CH4
Manfro et al (2011)
[5-85 wt]
PtAl2O3
160 - 280 autogenous highest H2 content (63-67) of gas phase at 230 degC
at 5-45wt glycerol
Oumlzguumlr and Uysal (2011)
31
The use of low temperature in the APR process results in the need for a highly active
catalysts such as Pt-based catalysts which are among the most widely studied Cortright et al
(2002) and Lehnert and Claus (2008) studied several Pt catalysts supported on Al2O3 Alumina-
supported Pt catalysts prepared from different Pt precursors ie platinum ethanolamine
platinum(II)-nitrate platinum sulfite and tetraammine platinum(II)-nitrate exhibited
comparable activities (~45 glycerol conversion) and selectivity towards H2 (~85) The
selectivity to H2 was found to increase from 78-95 with increasing catalyst particle size
ranging from 16-32 nm while the degree of conversion remained constant at ~20 The use
of a mixture of γ- θ- and δ-Al2O3 as support materials resulted in an increase in H2 production
as compared to pure γ-Al2O3 Luo et al (2008) examined the effect of Ptγ-Al2O3 with Pt
loading up to 12 wt at 180-220 degC and 114-25 bar on APR of 5 wt and 10 wt glycerol
solution At 5 wt ~65 glycerol conversion with ~45 yield of H2 was obtained at 220 degC
25 bar and the time on stream of 7 h The authors inferred that the H2 generation on the Pt
catalysts is accompanied by other parallel reactions such as dehydration hydrogenation and
methanation The Al2O3 supports of the used 09 wt Ptγ-Al2O3 catalysts (temperature not
specified) transformed into boehmite The formation of carbonaceous entities on the used
catalysts were the main cause for catalysts deactivation
Despite the high performance the high price of Pt-based catalysts usually limits their
use to small scale or laboratory applications Alternative transition metal catalysts have
therefore gained increasing interest Iriondo et al (2008) carried out APR of 1 wt glycerol at
225 degC 30 bar over Ni catalysts supported with Al2O3 using Ce La Mg or Zr as promoters It
was found that the Ni catalysts suffered from severe deactivation due to gradual Ni oxidation
while the La promoted catalysts provided the highest glycerol conversion of 37 mol (product
yield was not reported) the gaseous products included 32 mol H2 40 mol CO2 and 28
mol CH4 while the liquid products composed of 57 mol propylenglycol and 34 mol
32
ethylenglycol Wen et al (2008) studied the activity of various metal catalysts including Pt
Ni Co and Cu and their supports including SAPO-11 MgO H-USY zeolite Al2O3 SiO2 and
active carbon (AC) for APR of 10 wt glycerol at 230 degC 32 bar using a fixed-bed reactor
The activity of the four metal catalysts supported on Al2O3 decreased in the following order
Pt gt Cu gt Ni gt Co PtAl2O3 provided ~19 glycerol conversion and a gas containing ~70
mol H2 the overall yield rate of H2 production was 572 micromolmin-1119892119888119886119905minus1 PtMgO catalysts
increased the H2 yield and its rate of formation by absorption of formed CO2 by the support
~14 mol glycerol conversion with the H2 production rate of 432 micromolmin-1119892119888119886119905minus1 was
reported which is 25 lower than that obtained with PtAl2O3 In contrast an acidic zeolite H-
USY or neutral Al2O3 tended to increase the alkane formation
The gasification of crude glycerol has been applied on industrial scale since 2010 by
the BioMCN in The Netherlands with a production capacity of 200 kty (Blokland 2011) The
resulting syngas is reformed at high temperature and pressure in a conventional packed bed
methanol synthesis reactor to yield bio-methanol and the whole process is referred to as
glycerol-to-methanol (GTM) process (Van Bennekom et al 2012) The full carbon cycle of
the biodiesel production can be achieved as the bio-methanol may be re-used as a reactant in
the upstream transesterification process The syngas obtained may also be used as a reactant in
the FischerndashTropsch reaction for the production of fuels and chemicals (BTGgroup 2012)
Despite these the GTM process is uneconomical the price of bio-methanol (approx euro475-
525) is higher than that of petroleum-based methanol (approx euro200-250) (BioMCN 2013
Broeren 2013) The SCW gasification of crude glycerol as an alternative process for the
production of H2 or syngas would offer significant economic advantage over the current
gasification process since it does not require expensive drying step Drying and distillation
33
often make up ~50 of the total energy required for example in the production of ethanol
from corn grain (Peterson et al 2008)
223 Production of bio-chemicals by hydrothermal conversion of
biomass
The conversion of biomass into value-added chemicals represents one of the main
potential applications of hydrothermal technology The fermentation process currently the
main industrial process for the production of biomass-based chemicals has a number of
drawbacks such as the sensitivity of enzymes small space-time yields (Dusselier et al 2013)
and the inability to process mixed or complex feedstocks like lignocellulosic materials (Badger
2002) The flexibility and the potential of hydrothermal technology to overcome the drawbacks
of competing technologies was noted by the industry resulting in intensive research efforts over
the last two decades to find more economical cleaner and safer process
Starch and sugars are the main sources of biomass feedstock of current biochemical
production processes For instance the fermentation of corn sugars (or sugarcane or other
crops) is used for the industrial production of ethanol and lactic acid The sugars contained in
these feedstocks are easy to extract and convert into a wide variety of compounds including
the top twelve building block chemicals identified by the US Department of Energy (DoE)
see Figure 10 (Werpy and Petersen 2004) These building block chemicals may be used as
starting materials for production of a number of chemicals and polymers via catalytic routes
reviewed by ten Dam and Hanefeld (2011) and Dusselier et al (2013)
34
Figure 10 Top twelve building block chemicals identified by the US Department of Energy
(DoE) (Werpy and Petersen 2004)
Hydrothermal conversion of sugars has been studied at a wide range of temperatures
and pressures with or without a catalyst Table 6 summarises the research studies including
experimental conditions and results reported on hydrothermal conversion of various sugars to
chemicals
35
Table 6 The summary of research studies including experimental conditions and results on hydrothermal conversion of various sugars into
chemicals The percentage values are given in mol unless stated otherwise
Feedstock Temperature Pressure
Time
Catalysts Key products
Ref
Sucrose 250 ˚C 345 bar 25-100s H2SO4 glucose fructose glyceraldehyde furfural
dihydroxyacetone (DHA) lactic acid
5-(hydroxymethyl)furfural (HMF)
Antal Jr et al (1990)
Fructose 250 ˚C 345 bar 25-100 s H2SO4 HMF furfural pyruvaldehyde glucose formic
acid levulinic acid lactic acid glyceraldehyde
Antal Jr et al (1990)
glucose 300-400 ˚C 250-400 bar
002-2 s
- fructose glyceraldehyde DHA glycolaldehyde
pyruvaldehyde erythose 16-anhydro-D-glucose
Kabyemela et al (1999)
Fructose 300-400 ˚C 250-400 bar
002-2 s
- glyceraldehyde DHA erythrose Kabyemela et al (1999)
Glucose 250-350 ˚C 276 bar 5
min
NaOH lactic acid glycolic acid formic acid Calvo and Vallejo (2002)
cellobiose 350-400 ˚C 250-400 bar
001-054 s
- glucose erythrose glycolaldehyde Sasaki et al (2002)
D-glucose other
mono-saccharides
340 ˚C 275 bar 25-204 s HCl NaOH HMF glycolaldehyde glyceraldehyde formic acid
acetic acid lactic acid acrylic acid 2- furaldehyde
124-benzenetriol
Srokol et al (2004)
sucrose hexoses
trioses
300 ˚C 250 bar 10-180 s sulphate of
Co2+ Cu2+ Ni2+
Zn2+
lactic acid Bicker et al (2005)
D-fructose 200-320 ˚C autogeneous
75-180 s
HCl H2SO4
H3PO4 other
organic acids
HMF levulinic acid formic acid Salak Asghari and Yoshida
(2006)
36
The conversion of sugars under hydrothermal conditions typically occurs rapidly even
without a catalyst Kabyemela et al (1999) demonstrated that the decomposition of fructose
and glucose was almost complete within 2 s at 350 degC 250 bar while the carbon balance of
liquid products was above ~90 at short residence times but decreased at longer residence
times Despite the high conversion degree and conversion rate a mixture of several liquid
reaction products see Table 6 resulted in relatively low yields of individual compounds
As opposed to the non-catalysed reactions discussed above the catalysed reactions
generally provide higher yield of liquid products For instance the addition of H2SO4 in the
hydrothermal conversion (250 degC 345 bar) of fructose increased the yield of HMF from ~30
mol to ~55 mol (Antal Jr et al 1990) Takeuchi et al (2008) investigated the effect of
H3PO4 H2SO4 and HCl on the hydrothermal conversion of glucose at 250 degC autogenous
pressure and the reaction time of 1-10 min obtaining 40 C yield of HMF and ~10 C of
levulinic acid in the presence of H3PO4 at 250 degC in 5 min When HCl was used all HMF
rehydrated to form levulinic acid with the highest yield of ~55 C Yan et al (2010) employed
NaOH and Ca(OH)2 as catalysts in the transformation of glucose in sub-CW at 250-330 degC
273 bar finding lactic acid as the main product Similarly Esposito and Antonietti (2013)
examined the effect of NaOH NH4OH BaCl2 Ca(OH)2 Sr(OH)2 and NaOHBaCl2 on the
hydrothermal conversion of glucose at 220-250 degC autogenous pressure obtaining ~57 mol
yield of lactic acid at 250 degC using Ba(OH)2 at the base to glucose molar ratio of ~4 The
conversion of glucose under alkaline conditions has been long known to favour the formation
of lactic acid (Shaffer and Friedemann 1930) but Bicker et al (2005) demonstrated that the
sulphate of Co(II) Ni(II) Cu(II) and Zn(II) catalyse the conversion of various sugars to lactic
acid in sub-CW (300 degC 250 bar 10-180 s) 400 ppm ZnSO4 provided the lactic acid yields of
~42 wt and ~86 wt from sucrose and dihydroxyacetone respectively
37
Despite currently being an important feedstock for production of bio-chemicals the
sugars and starch containing materials form an integral part of human food chain resulting in
their high costs Using cheaper starting materials such as lignocellulose may reduce the cost
of bio-chemicals Lignocellulose naturally contained in wood and various fibrous plants is
abundant worldwide and outside of the human food chain The forests comprise approx 80
of the worldrsquos biomass (Badger 2002) hence the lignocellulosic biomass has been intensively
studied as a potential source of chemicals
Lignocellulosic materials are comprised of cellulose hemicellulose and lignin Due to
their polymeric and highly complex chemical structure the direct conversion of these materials
into chemicals requires a pre-treatment step to break down their structure into low molecular
weight sugars such as glucose (Mok and Antal 1992 Mok et al 1992 Adschiri et al 1993)
which can be further degraded into other valuable compounds by other conversion processes
such as fermentation and hydrogenation See Figure 11 for different treatment methods for
lignocellulosic materials and sugars
38
Figure 11 The treatment methods for lignocellulosic materials and sugars (adapted from E4tech et al (2015))
39
Mok and Antal (1992) investigated the non-catalysed hydrothermal treatment (200-230
degC 345 bar 0-15 min) of six woody and herbaceous biomass species containing mainly
hemicellulose (15-23) cellulose (30-45) and lignin (15-37) Between 40-60 of the
sample mass was solubilized the hemicellulose was completely solubilized resulting in 6-14
yield of monomeric sugar whereas the cellulose and lignin were only partially solubilized
Adschiri et al (1993) demonstrated that the yield of glucose may be increased by pre-treating
the biomass in SCW at 400 degC 250 bar where ~100 mol conversion of cellulose with ~34
mol yield of glucose was obtained within 15 s Despite being effective the process requires
high pressure and temperature On the other hand Mok et al (1992) reported ~71 mol yield
of glucose after a pre-treatment of cellulose in hot-compressed water at 215 degC 345 bar using
H2SO4 as a catalyst In fact acid-base catalysts are often preferred in hydrothermal treatment
of lignocellusolics making the process not only rapid but also highly selective However the
corrosiveness of the catalysts combined with that from the self-ionization of water leads to
severe equipment corrosion Onda et al (2008) examined the potential of several less corrosive
solid acid catalysts such as H-zeolites activated carbon (AC) SiO2 γ-Al2O3 AC-SO3H
sulfated zirconia and amberlyst 15 for the hydrothermal conversion of cellulose and starch at
150 degC autogenous pressure and a reaction time of 24 h AC-SO3H gave high yields of glucose
~405 C from cellulose and ge90 C from starch possibly due to its (i) high hydrothermal
stability and (ii) the presence of strong acid sites (-SO3H) and the hydrophobic surface Both
sulfated zirconia and amberlyst 15 catalysts gave relatively high yields of glucose with high
selectivity to water soluble by-products but were not stable under the conditions tested The H-
zeolite catalysts gave only slightly higher yields than the non-catalysed reactions whilst the γ-
Al2O3 and SiO2 were completely inactive
In addition to mono-sugars the hydrothermal conversion of lignocellulosic feedstocks
yields chemicals with high market value such as 5-(hydroxymethyl)furfural (HMF)
40
glyceraldehyde dihydroxyacetone (DHA) pyruvaldehyde and organic acids (see Table 7) all
of which are the decomposition or transformation products of mono-sugars A single-step
catalytic direct conversion of lignocellulosic biomass into chemicals gained significant interest
in recent years The studies published on this topic including the experimental conditions and
results are summarised in Table 7
41
Table 7 Subcritical and supercritical water conversion of various lignocellulosic biomass into chemicals The percentage values are given in mol
unless stated otherwise Note S = selectivity Y = yield
Feedstock Temperature Pressure
Time
Catalysts Key products Ref
cellulose 215 degC 345 bar 1-60 min H2SO4 glucose HMF levoglucosan cellobiose Mok et al (1992)
cellulose 200-400 degC 250 bar 5-
300 s
- glucose fructose erythrose glyceraldehyde organic
acids pyruvaldehyde DHA glycolaldehyde 16-
anhydro-D-glucose
Adschiri et al (1993)
microcrystalline
cellulose
320-400 degC 250 bar 005
to 10 s
- at 320 ordmC and 350 ordmC the main products were organic
acids with Y = 53 and 606 respectively
at 400 ordmC main product was oligomers with C gt6 with
Y = 51
Sasaki et al (2000)
rabbit food (municipal
solid waste)
200-350 degC na 20-150 s Y (glucose) = 33 Y (lactic acid) = 32 Y (acetic
acid) = 26
Goto et al (2004)
cherry (hard wood)
cypress (soft wood)
280 degC na 15 min K2CO3 50 wt total oil yields 10-20 wt gases ~24-43 wt
water soluble compounds (mainly acetic acid)
Bhaskar et al (2008)
cellulose starch
glucose
270-400 degC auto-
geneous 30-180 s
NaOH Ca(OH)2 Y (lactic acid) = 20-27 Yan et al (2010)
Cellulose 180-220 degC na 10-200
min
alkali alkaline
earth AlCl3 and
transition metal
chlorides
CrCl3 was exceptionally effective Y (levulinic acid) =
67
Peng et al (2010)
Cellulose 300 degC autogenous 5
min
Ni Zn activated
C in NaOH
Y (lactic acid) = 42 Zhang et al (2011)
Cellulose 325 ˚C autogenous 1 min CuO in NaOH Y (acetic glycolic lactic formic acids) = 43 Wang et al (2014)
42
224 Production of bio-chemicals by hydrothermal conversion of
glycerol
The majority of studies performed on hydrothermal conversion of glycerol tested the
effect of acid or base catalysts and oxidation agents such as H2O2 The acid catalysed
hydrothermolysis of glycerol resulted in dehydration and formation of acrolein acetaldehyde
and acetol (1-hydroxyacetone) (Watanabe et al 2007) while the bases induced
dehydrogenation usually followed by dehydration and re-hydration forming lactic acid one of
the oxidation products of glycerol (Kishida et al 2005 Ramirez-Lopez et al 2010) The
hydrothermal oxidation of glycerol led to a variety of decomposition products including
glyceric acid dihydroxyacetone lactic acid acetic acid formic acid etc (Katryniok et al
2011) Figure 12 shows the reaction pathways of glycerol and its conversion products under
hydrothermal conditions
43
Figure 12 The reaction pathways of glycerol under hydrothermal conditions (adapted from
Zhou et al (2008)) Katryniok et al (2011) Akizuki and Oshima (2012))
The current productionprocess price and applications of common reaction products of
glycerol (highlighted in green in Figure 12) are summarised in Table 8 and discussed in detail
in this section The hydrothermal conversion of glycerol in the presence of acids bases and
oxidants are discussed in detail in Chapters 2242 2243 and 2244 respectively
44
Table 8 The current productionprocess price and applications of selected reaction products
of glycerol
Compound Current production
process
Production
[tonnesyr]
Price
US $tonnes
Examples of applications
acrolein oxidation of propene 500k1 1850-20005 acrylic acid DL-methionine
acetaldehyde oxidation of ethylene 1280k2 4500-65005 pyridine (derivatives)
pentaerythritol
crotonaldehyde
lactic acid fermentation of
carbohydrates
400k3 1300-160056 Food amp food related
applications fibres green
solvents poly(lactic acid)
DHA fermentation of
glycerol
2k4 150000-16500057 Cosmetic industry
1(Patocka and Kuca 2014) 2(Eckert et al 2006) 3(Dusselier et al 2013) 4(Pagliaro et al 2007)
5(wwwalibabacom 2016) 6(L Nattrass 2010) 7(Katryniok et al 2011)
(i) Acrolein
Acrolein (prop-2-enal) the simplest unsaturated aldehyde is a colourless volatile
toxic and lacrimatory liquid with a strong odour (Etzkorn 2000) It is an important intermediate
commonly used in the chemical and agricultural industries Due to its toxicity and instability
acrolein is usually produced as an immediate which is immediately converted into other
chemicals mostly acrylic acid the key precursor of polyacrylic acid (Katryniok et al 2009)
Much smaller proportion of acrolein is isolated and used primarily for the synthesis of DL-
methionine (see Figure 13) an important amino acid used predominantly in meat production
to accelerate the animal growth (Etzkorn 2000 Katryniok et al 2009) The ever increasing
global demand for DL-methionine is currently satisfied by its industrial synthesis from
acrolein see Figure 13
45
Figure 13 The synthesis of DL-methionine from acrolein (adapted from Katryniok et al
(2009))
Acrolein is industrially manufactured by oxidation of propene using multicomponent
BiMoOx based catalysts (Arntz et al 2000) see Figure 14A Approximately half a million tons
of acrolein are produced annually however the dependence of this process on petrochemicals
makes it unsustainable (Patocka and Kuca 2014) Therefore the preparation of acrolein from
low cost crude glycerol is economically and environmentally more attractive see Figure 14B
Figure 14 The formation of acrolein by (A) selective oxidation of propene (B) dehydration of
glycerol
The acid catalysts are known to selectively catalyse the conversion of glycerol to
acrolein (Katryniok et al 2013) However the use of strong mineral acids such as H2SO4 to
catalyse the dehydration of glycerol to acrolein represents an issue which stimulated significant
research efforts in recent years on development of an acid-free process Various types of
46
catalysts including zeolites silica-alumina materials heteropoly acids metal oxides etc were
investigated see Chapter 2242
(ii) Acetaldehyde
Acetaldehyde is an important platform chemical and a dehydration product of glycerol
with a market price of US $ 4500-6500tonnes (wwwalibabacom 2016) It is an important
intermediate of pyridine pyridine derivatives pentaerythritol and crotonaldehyde production
(Sowin and Melcher 2001) It is a precursor for the synthesis of acetals such as 11-
diethoxyethane see Figure 15 used in the fragrance industry and a diesel additive increasing
its cetane number promoting complete combustion thus reducing the emissions of CO
unburnt-hydrocarbons and other smog-causing gases (Frusteri et al 2007) The preparation of
11-diethoxyethane from acetaldehyde and ethanol received significant attention since both
reactants may be obtained from renewable sources (Andrade et al 1986 Kaufhold and El-
Chahawi 1996 Capeletti et al 2000 Gomez et al 2004)
Figure 15 The production of 11-diethoxyethane from acetaldehyde and ethanol
The world-wide production of acetaldehyde estimated to be 128 Mtyear (Eckert et
al 2006) relies mainly on the Wacker-Hoechst direct oxidation of ethylene ie a non-
47
renewable chemical derived from petroleum The production of acetaldehyde from bioethanol
is an alternative sustainable process (Murray et al 1990 Moroz et al 2000) however the use
of crude glycerol a by-product of the biodiesel industry instead of bioethanol for the
production of acetaldehyde would potentially further improve the economy of the biodiesel
industry due to its higher availability and lower cost
(iii) Lactic acid
Lactic acid (LA) (2-hydroxypropanoic acid) is an important commodity chemical with
a price of US $1300-1600tonne ie at least six times higher than that of crude glycerol (L
Nattrass 2010) LA has been long used by a number of industries mainly in the food and food-
related applications (Datta and Henry 2006) Its worldwide production in 1995 reached 50000
tonnes and ~300000 to 400000 tonnes in 2014 (Datta and Tsai 1997 L Nattrass 2010
Dusselier et al 2013) Such increase may be accounted to its newly emerging uses as a benign
chemical feedstock for producing fibres green solvents oxygenated chemicals and more
importantly biodegradable poly(lactic acid) (PLA) PLA found a multitude of uses in the
medical industry for instance as biodegradable surgical sutures orthopaedic implants as well
as in controlled drug release (Okada et al 1994 Middleton and Tipton 2000 Narayanan
2004) PLA is one of the most commercially used biodegradable polymers with a market value
of US $2200tonne (Madhavan Nampoothiri et al 2010)
Currently LA is commercially produced via fermentation of carbohydrates such as
corn This process however suffers from several limitations as it requires a long reaction time
has a low efficiency and generates significant amount of waste gypsum resulting in high
operating cost (Dusselier et al 2013 Mazumdar et al 2013) Catalytic conversion of glycerol
to LA was demonstrated to be a promising alternative (see Chapter 2243)
48
(iv) Dihydroxyacetone
Dihydroxyacetone (DHA) (13-dihydroxypropanone) is a simple carbohydrate with
high market value ~150 US $kg used mainly in the cosmetic industry as an active ingredient
in sunless tanners (Katryniok et al 2011) It may also be used as a precursor for synthesis of
lactic acid High yield of lactic acid was reported using DHA 86 wt (Bicker et al 2005) 71
mol (West et al 2010)
DHA is industrially produced via glycerol fermentation Despite providing high
selectivity to DHA this process has several drawbacks such as low productivity and high
production cost A range of contributing factors exists for instance high glycerol
concentrations and the produced DHA inhibit microorganism growth The impurities in crude
glycerol if used may have a similar effect (Szymanowska-Powalowska and Bialas 2014)
These limitations prevent the fermentation process from reaching the desired effectivity and
meeting the market demands for DHA A more efficient production process of DHA from
glycerol such as catalytic hydrothermal conversion which is discussed in detail in Chapter
2244 is desirable
2241 Non-catalytic hydrothermal conversion of glycerol
Hydrothermal dehydration of glycerol to produce acrolein has been investigated as an
alternative to gas-phase dehydration claimed to be an economically feasible process but has
not proven competitive with the production route from petrochemicals The potential of
hydrothermal dehydration of glycerol not only for the preparation of acrolein but also for other
dehydration products such as acetol and acetaldehyde were demonstrated by a number of
studies
49
Antal et al (1985) were among the first to study the hydrothermal dehydration of
glycerol (at 360 degC and 500 degC 350 bar) detecting acetaldehyde acrolein H2 CO CO2 CH4
C2H2 and C2H6 as products The selectivity towards acetaldehyde was generally higher than
towards acrolein at 360 degC and much higher at 500 degC The non-catalysed glycerol degradation
was examined under at 200-400 degC 300 bar residence time of 20-60 min by Qadariyah et al
(2011) who observed ~100 mol glycerol conversion in a batch reactor at ge250 degC 300 bar
allyl alcohol was found only in SCW with a yield of 97 mol at 400 degC in 30 min
Acetaldehyde was detected in sub-CW region at 250 degC while acrolein was found in both sub-
and supercritical regions The dehydration of glycerol in a continuous flow system was studied
by Buumlhler et al (2002) who on the contrary observed acetaldehyde as the major product at
451 degC 450 bar with the yield increasing with the reaction time Table 9 summarises the works
reported on the non-catalysed hydrothermal conversion of glycerol
50
Table 9 The summary of experimental works including experimental conditions and results on non-catalysed hydrothermal conversion of
glycerol The percentage values are given in mol unless stated otherwise
Reactor Type Temperature Pressure Time Products Ref
- 360 degC 500 degC 350 bar 90 s X = 32 products acrolein acetaldehyde H2 CO CO2 CH4
C2H4 C2H6
Antal et al (1985)
flow 349-475 degC 250 350 450 bar
32-165 s
X = 04-31 products MeOH EtOH acetaldehyde
propionaldehyde acrolein allyl alcohol formaldehyde CO CO2
H2
Buumlhler et al (2002)
batch 300 degC saturated pressure 10-
60 min
X = 2-5 products acrolein acetaldehyde allyl alcohol
formaldehyde hydroxyacetone
Watanabe et al
(2007)
batch 200-400 degC 300 bar 20-60 min X = 999 products allyl alcohol (Y = 967) acetaldehyde
acrolein
Qadariyah et al
(2011)
Note X = glycerol conversion S = selectivity Y = yield
51
2242 Hydrothermal conversion of glycerol catalysed by acids
The non-catalysed dehydration of glycerol yields primarily acetaldehyde Buumlhler et al
(2002) reported a maximum of 31 mol dehydration of 019-057 M glycerol in water at 300-
474 degC 250 350 and 450 bar residence time of 32-165 s Acetaldehyde with the highest
selectivity of 81 mol was found to be the main liquid product under all conditions studied
This was previously observed by Antal et al (1985) who investigated the hydrothermal
decomposition of glycerol at 360 degC and 500 degC 350 bar finding that the selectivity towards
acetaldehyde was generally higher than towards acrolein at 360 degC and much higher at 500 degC
It was concluded that acetaldehyde may formed through three main routes see Figure 15 of
which the Route 3 was dominant in the absence of a catalyst On the other hand the addition
of NaHSO4 increased the acrolein formation at 360 degC whereas at 500 degC the ratio of
acetaldehyde to acrolein selectivities was not significantly affected by NaHSO4 possibly due
to the fact that the free radical chemistry may play the dominant role in glycerol decomposition
at higher temperatures
52
Figure 16 The dehydration reactions of glycerol under hydrothermal conditions (adapted
from Antal et al (1985))
Most of later studies focussed on the effect of catalysts to decrease the reaction
temperature as well as to increase the yield of desired products Table 10 summarises the
experimental works including experimental conditions and results on acid-catalysed
hydrothermal conversion of glycerol
53
Table 10 The summary of experimental works including experimental conditions and results on acid-catalysed hydrothermal conversion of
glycerol The percentage values are given in mol unless stated otherwise Note X = glycerol conversion S = selectivity Y = yield
Catalyst Reactor
Type
Temp Press Time Results
Ref
NaHSO4 - 360 degC 500 degC 350 bar 90 s S (acrolein) = 70 S (acetaldehyde) = 35 Antal et al (1985)
H2SO4 flow 300-350 degC 345 bar 16-39 s X = 10-55 products acrolein
acetaldehyde CO CO2 C2H4
Ramayya et al (1987)
H2SO4 flow 300-400 degC 250 300 345 bar
5-83 s
X = 90 S (acrolein) = ~80 Watanabe et al (2007)
ZnSO4 flow 300-390 degC 250-340 bar 10-
60 s
X = 50 S (acrolein) = 76 Ott et al (2006)
RuZrO2 flow 510-550 degC 350 bar 2-10 s X = 100 products acetaldehyde acetic
acid hydroxyacetone acrolein H2 COCO2
May et al (2010)
TiO2 WO3TiO2 flow 400 degC 330 bar WF (01-8)
x103 kg-cat sm-3
X = ~100 products acrolein
acetaldehyde hydroxyacetone allyl alcohol
Akizuki and Oshima (2012)
TiO2 WO3TiO2 flow 400 degC 250-410 bar WF
(005- 8) x103 kg-cat sm-3
acrolein acetaldehyde hydroxyacetone
lactic acid
Akizuki and Oshima (2013)
H-Beta H-MFI zeolites Na-
Beta Na-MFI
flow 150-210 degC na flow rate 12
mlmiddoth-1
X = ~60 Y (acrolein) = 29 Lin et al (2013)
H-ZSM-5 H-Mordenite flow 180-340 ordmC 70 bar
LHSV = 2 h-1
acrolein Neher et al (1995)
H-Y H-Beta Mor SBA-15
ZSM-23
batch 250 ordmC 70 bar 10 h X = ~90 S (acrolein) = 100 de Oliveira et al (2011)
WF was used as an indicator of reaction time it is defined as the amount of catalyst divided by volumetric flow rate
54
Ramayya et al (1987) reported a high conversion (up to 55 mol) and high yield of
acrolein (up to 86 mol) from the H2SO4 catalysed hydrothermal conversion of glycerol at
300-350 degC 345 bar residence time of 16-39 s at the glycerol to H2SO4 molar ratio ~100 On
the other hand Antal et al (1990) found that only 1 of total glycerol in 01 M glycerol - 001
M H2SO4 solution converted at 250 degC 345 bar concluding that the optimum conditions for the
formation of acrolein include the temperatures above 250 degC and the use of a mineral acid
Identical conclusion was made by Watanabe et al (2007) who observed ~90 mol glycerol
conversion with ~80 mol acrolein selectivity in the H2SO4-catalysed degradation of glycerol
at 400 degC 345 bar
Despite the high conversions and acrolein yields reported the use of mineral acids for
the hydrothermal conversion of glycerol raised concerns over the corrosive effect on the
equipment used This issue may be alleviated by the use of alternative less corrosive catalysts
such as
metal salts
metal oxides and
zeolites
Metal salts
Ott et al (2006) found that ZnSO4 had a competitive catalytic effect to the more
corrosive H2SO4 50 mol glycerol conversion with a maximum acrolein selectivity of 75
mol at 300-390 degC 250-340 bar 10 to 60 s was obtained
55
Metal oxides
May et al (2010) reported the formation of dehydration products of glycerol during the
1 RuZrO2 catalysed gasification of glycerol in SCW at 550 degC 350 bar almost complete
glycerol conversion with ~55 mol selectivity to liquid products were achieved in 4 s Akizuki
and Oshima (2012) investigated the effect of WO3TiO2 on the conversion of glycerol in SCW
(400 degC 33 MPa) using a continuous flow reactor The concentration of acrolein the main
product was increasing with the WO3 concentration reaching a maximum yield of 53 at
100 glycerol conversion Other minor products included acetaldehyde propionic acid
hydroxyacetone lactic acid allyl alcohol propionaldehyde acetone acetic acid and acrylic
acid Despite the promising results of these metal oxide catalysts the high cost for working
under supercritical conditions may prevent this process from future commercialization
Zeolites
Due to the low thermal stability of zeolites in sub-CW amp SCW zeolite catalysts were
commonly used in gas phase- as opposed to liquid phase- glycerol conversion (Corma et al
2008 Kim et al 2010 Carrico et al 2013)
Neher et al (1995) patented a process for the production of acrolein by dehydration of
glycerol in both liquid- and gas phase using acidic solid catalysts such as H-ZSM-5 Na-ZSM-
5 γ-Al2O3 and Mordenite The liquid-phase dehydration of 10 wt glycerol was performed in
a fixed bed reactor with the liquid hourly space velocity (LHSV) of 2 h-1 at 180-340 deg C 70
bar H-ZSM-5 catalysts (SiO2Al2O3 = 60 and 28) showed the highest activity with ~16-20
conversion and ~71-75 selectivity to acrolein The catalysts maintained their activity even
after 50 h of operation Na-zeolite was inactive at all while γ-Al2O3 gave insufficient
selectivity
56
de Oliveira et al (2011) reported a high yield of acrolein from liquid-phase dehydration
of glycerol under hydrothermal conditions (250 degC 70 bar) using several protonic zeolites (H-
zeolites) of which protonic Y zeolite (H-Y) and protonic Mordenite zeolite (H-MOR) provided
the highest conversion (~90 mol) and selectivity (100 mol) to acrolein The liquid phase
dehydration of glycerol at 150-210 degC was studied by Lin et al (2013) The process was
investigated in a fixed bed reactor using microporous H-Beta zeolite (SiAl ratio = 158) and
H-MFI (SiAl = 174) with different concentrations of Broslashnsted acidic sites introduced by
partial replacement of the balancing H+ cation of the zeolites with non-active Na+ The acrolein
yield of 29 mol was achieved at 60 mol glycerol conversion at 200 degC demonstrating the
potential of zeolites as catalyst for hydrothermal dehydration of glycerol
2243 Hydrothermal conversion of glycerol catalysed by bases
The majority of studies on base-catalysed hydrothermal conversion of glycerol focusses
on optimising the conditions for maximum yield of lactic acid (Kishida et al 2005 Shen et
al 2009 Roy et al 2011) Table 11 summarises the studies including experimental
conditions and results on base-catalysed hydrothermal conversion of glycerol
57
Table 11 The summary of studies including experimental conditions and results on base-catalysed hydrothermal conversion of glycerol The
percentage values are given in mol unless stated otherwise
Glycerol Catalyst Reactor Type Temperature Pressure
Time
Conversion
[]
Yield of
LA2 [] [LA]3 gL Ref
033 M glycerol 125 M
NaOH
Batch 300 degC na 90 min 100 90 267 Kishida et al (2005)
033 M glycerol 125 M
KOH
Batch 300 degC ~90 bar1 60 min 100 90 267 Shen et al (2009)
25 M glycerol 275 M
NaOH
Batch 280 degC na 90 min 100 845 1903 Ramirez-Lopez et al
(2010)
011 M glycerol PtC
RuC in 08 M CaO NaOH
Semi-
batchwise
200 degC 40 bar H2 5 h 100 58 57 Maris and Davis (2007)
01 M glycerol PtCaCO3
in 01 M H3BO3
Batch 200 degC 40 bar H2 18 h 45 54 49 ten Dam et al (2011)
03 M glycerol CuO2 in
045 M NaOH
Batch 200 degC 14 bar N2 6 h 978 76 205 Roy et al (2011)
054 M glycerol IrC RhC
in 054 M NaOH
Batch 180 degC H2 or He na ~50 - Auneau et al (2012)
033 M glycerol
01 molL-1 CaO suspension
Batch 290 degC na 150 min 978 408 121 Chen et al (2014)
1 saturated vapour pressure 2 LA=lactic acid 3 concentration
58
Kishida et al (2005) achieved a complete conversion of glycerol with 90 mol yield
of lactate at 300 degC 90 min utilizing NaOH as a catalyst at the NaOH to glycerol molar ratio
of 379 Large amount of H2 was also detected Figure 17 shows the proposed mechanism of
the reaction between glycerol and NaOH forming sodium lactate and H2 under hydrothermal
conditions (Kishida et al 2005) An almost identical mechanism was proposed by Ramirez-
Lopez et al (2010) Under hydrothermal conditions glycerol first reacts with NaOH to form
sodium glyceroxide which loses a proton (becoming a hydride ion H-) to yield glyceraldehyde
Glyceraldehyde then undergoes hydrogen abstraction at the C2 atom by the released hydride
ion (H-) followed by OH elimination to give 2-hydroxypropenal By keto-enol tautomerization
2-hydroxypropenal transforms into pyruvaldehyde which is then converted into lactic acid via
benzilic acid rearrangement or internal Cannizaro reaction
Figure 17 The proposed reaction mechanism of the reaction between glycerol and NaOH
forming sodium lactate and H2 (Kishida et al 2005 Ramirez-Lopez et al 2010)
59
Shen et al (2009) investigated the effect of various alkaline catalysts on the
hydrothermal conversion of glycerol to lactic acid NaOH KOH LiOH Ba(OH)2 Sr(OH)2
Ca(OH)2 Mg(OH)2 and Al(OH)3 were investigated at 300 degC 90 bar 90 min at the base to
glycerol molar ratio of 379 All catalysts were found to be active except for Al(OH)3 High
yields of lactate salts (90 mol) were obtained using either KOH or NaOH which is in
accordance with previous observation made by Kishida et al (2005) The formation of lactate
in the hydrothermal conversion of glycerol depended not only on the hydroxide concentration
but also on the type of metal ions of catalysts For instance KOH was a better catalyst than
NaOH requiring lower concentration and a shorter reaction time to obtain a comparable lactate
yield The effectivity of catalysts decreased in the following order KOH gt NaOH gt LiOH gt
Ba(OH)2 gt Sr(OH)2 gt Ca(OH)2 gt Mg(OH)2
Aiming to develop an economically feasible industrial process for manufacturing of
lactic acid from glycerol Ramirez-Lopez et al (2010) investigated an alkaline hydrothermal
conversion of pure glycerol at glycerol concentration almost 10 times higher than that used by
Kishida et al (2005) and Shen et al (2009) The experiments were carried out at 250-290 degC
30-250 min and NaOH to glycerol molar ratio of 11-175 obtaining ~85 mol lactate yield at
280 degC within 90 min at the NaOH to glycerol molar ratio of 11 Identical yield was obtained
when crude glycerol was used The main by-product was sodium carbonate while other by-
products included sodium acrylate oxalate and formate
To avoid the high corrosiveness and problematic downstream separation associated
with the use of homogeneous alkali catalysts the potential of solid basic catalysts for the
hydrothermal conversion of glycerol to lactic acid was examined by Chen et al (2014) who
tested CaO Ca(OH)2 hydrotalcite dehydrated hydrotalcite and alumina loaded with 35 wt
KNO3 as catalysts The water content of glycerol used was 4-12 wt and the experiments were
carried out at 300 degC for 90 min with the catalyst to glycerol molar ratio of 379 A glycerol
60
conversion of 978 mol with 408 mol yield lactic acid was obtained on CaO at the base to
glycerol molar ratio of 03 290 degC 150 min Similar yield was obtained using crude glycerol
It was also found that the water content of glycerol higher than 10 wt decreased the yield of
lactic acid due to the formation of weakly basic Ca(OH)2
The addition of noble or transition metal catalysts is known to allow the alkaline
hydrothermal reactions to be performed at lower temperature Maris and Davis (2007) used
activated carbon-supported Ru and Pt catalysts in the aqueous-phase hydrogenolysis of
glycerol at 200 degC 40 bar H2 RuC showed better performance than PtC in the production of
ethylene glycol and propylene glycol In contrast the addition of bases (NaOH and CaO) led
to a significant lactate formation the extent of which was greater in the presence of Pt ~58
mol after 5 h as opposed to Ru ten Dam et al (2011) tested 19 commercially available
catalysts for the hydrogenolysis of glycerol to 12-propanediol at 150-200 degC 40 bar H2 The
PtCaCO3 was found to be active while the PtCaCO3 in the presence of H3BO3 decreased the
selectivity to 12-propanediol but increased the selectivity to lactic acid The role of H3BO3
was unclear It was possible to recycle PtCaCO3 without losing the selectivity although certain
degree of deactivation was observed Auneau et al (2012) investigated the carbon supported
Ir and Rh catalysed hydrothermal treatment of 5 wt glycerol solution at 160-180 degC under
He or H2 atmosphere and the glycerol to NaOH molar ratio of 1 A high yield of lactate ~50
mol was reported at 180 degC on the IrC catalyst under He atmosphere Roy et al (2011)
employed Cu2O in the conversion of glycerol at 200 degC in the presence of NaOH receiving 76
mol lactate yield after 6 h In addition to the use of low temperature unlike aqueous phase
hydrogenolysis the process required neither O2 nor H2 Also Cu2O showed good stability even
after several cycles of use However the concentration of feed glycerol (01-03M) was low
The reaction pathway of glycerol conversion to lactate under alkaline hydrothermal conditions
in the presence of metal-based catalysts was proposed to lead through the dehydrogenation of
61
glycerol to glyceraldehyde which is then induced by alkali to yield pyruvaldehyde
Pyruvaldehyde undergoes benzilic acid rearrangement to form lactate ion (Maris and Davis
2007 Auneau et al 2012) see Figure 18
Figure 18 The proposed conversion mechanism of glycerol to lactate under alkaline
hydrothermal conditions in the presence of metal-based catalysts Maris and Davis (2007)
Auneau et al (2012)
A batch hydrothermal electrolysis employing an Ir plate anode was reported by Yuksel
et al (2009) to be a promising method for the transformation of glycerol ~84 mol conversion
with 347 mol of lactate yield as well as H2 gas was obtained after the degradation of glycerol
at 280 degC in the presence of NaOH in 90 min A conversion exceeding 90 mol was later
reported (Yuksel et al 2010)
The performances of Ru Pt Au Rh Ir and Cu under alkali conditions at 180-200 degC
were competitive to that of alkaline metals at 280-300 degC (Maris and Davis 2007 Shen et al
2010 Roy et al 2011 Auneau et al 2012) This shows the potential of alkali hydrothermal
method as a process for preparing lactic acid from glycerol but the resulting lactate salts need
to be converted into free lactic acid by neutralizing with a strong acid (such as H2SO4)
generating metal salt wastes
62
2244 Hydrothermal conversion of glycerol with H2O2
Hydrothermal oxidation is a thermal oxidation process operating in hot-compressed
water using oxygen as an oxidizer It is also known as wet oxidation (WO) or wet air oxidation
(WAO) when air is used It has been commercialised since 1954 as an effective treatment
process for high water content organic waste not suitable for incineration (Zimmermann 1954)
Occasionally the oxidation of organics to CO2 and water is incomplete resulting in valuable
organic compounds such as acetic acid formic acid glycolic acid and lactic acid (Meyer et al
1995 Jin et al 2001 Jin et al 2001 Calvo and Vallejo 2002) and thus the hydrothermal
oxidation has gained an interest as a process for converting biomass into chemicals (Quitain et
al 2002 Jin et al 2005 Jin et al 2006)
Our literature survey reveals that the research on hydrothermal oxidation of glycerol
with an oxidant is most active in China The summary of works published is listed in Table 12
63
Table 12 Hydrothermal oxidation of various biomass into chemicals Percentages present are in mol unless stated otherwise
Feedstock Temperature
Pressure Time
Catalysts Key products
Ref
045 M Crude
glycerol
150-450 degC saturated
vapour pressure with air
or O2 or H2O2
with or without NaOH Y (formic acid) = 31 co-products oxalic
acid Y (formic acid) = ~35 when NaOH
was added
Zhang et al (2013) Zhang(a)
et al (2014)
2 mmol glycerol 120-180 degC 5-15 bar
O2 1 h 40 mM FeCl3
Ru(OH)4 on
reduced graphite oxide
Y (formic acid) = 60 (carbon basis) Xu et al (2014)
1-90 wt
glycerol
150 degC 20-40 bar O2 3h vanadium-substituted
phosphomolybdic acids
(H3+nPVnMo12-nO40)
X = 95wt Y (formic acid) = ~36 wt
other products acetic acid and
formaldehyde
Zhang(b) et al (2014) Zhang
(2015)
glycerol to Ru mol ratio = 1000 Note X = glycerol conversion S = selectivity Y = yield
64
Zhang et al (2013) studied the hydrothermal oxidation of 045 M glycerol with H2O2
at 250 degC autogenic pressure (170 bar) 16 s The highest formic acid yield of 31 mol was
obtained at H2O2 to glycerol molar ratio of 24 while the addition of NaOH expected to prevent
the oxidative decomposition of formic acid provided only ~35 mol yield It was proposed
that glycerol under oxidative hydrothermal conditions may oxidise to yield formic and oxalic
acids of which the latter possibly inhibited further decomposition of formic acid It was found
that the slightly higher yield of formic acid obtained under alkaline conditions may be due to
the initial formation of lactic acid and its subsequent decomposition to formic acid The
addition of NaOH did not inhibit the decomposition of formic acid possibly explaining the
similar yields obtained with or without NaOH The work was patented in 2014 (Zhang(a) et
al 2014)
Zhang(b) et al (2014) and Zhang (2015) reported that glycerol was selectively
converted into formic acid by O2 under hydrothermal conditions using vanadium-substituted
phosphomolybdic acids (H3+nPVnMo12-nO40) as catalysts The experiments were conducted in
a batch reactor at 150 degC 20-40 bar O2 for 3 h using glycerol solutions of different
concentrations ranging from 1-90 wt A complete conversion of glycerol was achieved for 1
wt solution and the main liquid products were formic acid acetic acid and formaldehyde with
the selectivity of ~51 wt ~8 wt and ~7 wt respectively while CO2 was the only product
detected in the gas phase It was found that higher concentration of glycerol resulted in lower
degree of conversion but concurrently higher absolute yield of formic acid and acetic acid
Increasing the vanadium content of the phosphovanadomolybdic catalyst also enhanced the
conversion efficiency of glycerol to formic acid The highest absolute yield of formic acid (~36
wt) was obtained from 50 wt glycerol at 150 degC 20 bar O2 for 3 h using H6PV3Mo9O40
This method was patented in 2015 (Han and Zhang 2015)
65
Xu et al (2014) reported high yield of formic acid (~60 mol) from the hydrothermal
oxidation of glycerol (120-180 degC 5-15 bar O2 1 h) catalysed by Ru(OH)4 doped on reduced
graphite oxide (r-GO) in the presence of Lewis acidic FeCl3 Both Ru(OH)4r-GO and FeCl3
could individually catalyse the glycerol oxidation giving formic acid as the main liquid product
albeit low glycerol conversion (lt20 ) Other Lewis acids tested included AlCl3 CrCl3 and
ZnCl2 each of which exhibited insignificant activity in the oxidation of glycerol
In contrast to alkaline hydrothermal reactions wet oxidation of glycerol directly
produces organic acids such as lactic acid acetic acid and formic acid in their free acid forms
However further research are still required to develop a commercially feasible processes
23 Zeolite-based catalysts
231 Zeolites and their Broslashnsted and Lewis acidity
Zeolites are commonly used as both catalyst supporting materials and solid acid
catalysts in a broad range of commercial processes including fluid catalytic cracking (FCC)
due to their high catalytic performances high surface area high thermal stability high ion
exchange capacity and low cost (Weitkamp 2000) Zeolites are crystalline microporous
aluminosilicates with 3D framework formed by tetrahedral [SiO4] and [AlO4] units linked to
each other by sharing oxygen atoms at their vertices see Figure 19
66
Figure 19 Broslashnsted and Lewis acid sites on zeolite adapted from Kondo et al (2010) and
Almutairi (2013)
Zeolites may act as both Lewis and Broslashnsted acids4 The presence of Al atoms in the
zeolite structure causes negative charges on the zeolite framework which have to be
compensated by cations to maintain the structural stability A wide range of cations
determining the extent of BroslashnstedLewis acidity of zeolites may be used For instance H+ as
a charge balancing-cation provides the zeolites with Broslashnsted acidity while (AlO)+ or metal
cations with Lewis acidity (Almutairi 2013) Since the Al atoms are the source of protons the
Broslashnsted acidity increases with the increasing AlSi ratio On the other hand the Lewis acidity
of zeolites may be controlled by the nature of foreign ions introduced by an ion-exchange
reaction replacing the balancing cations andor the aluminium atoms within the frameworks by
various multivalent metal cations such as Sn4+ Ti4+ and Zr4+ (Hammond et al 2012 Gounder
and Davis 2013)
4 A Lewis acid is a compound with the ability to accept lone pairs of electrons due to having one or more empty
orbitals while Broslashnsted acid is a compound capable of donating protons (H+)
67
232 Stability of zeolites under hydrothermal conditions
Over the last two decades zeolites have gained increasing interest as solid acid catalysts
and catalyst supports for the conversion of biomass into fuels and chemicals (Triantafillidis et
al 2000 Taarning et al 2011) Their high surface acidity makes them one of the most
promising alternatives to highly-corrosive acid additives such as H2SO4 in the conversion of
glycerol carbohydrates and lignocellulosic materials (Klinowski 1984 Corma et al 2008
Katryniok et al 2013) However the stability of zeolites is limited when used in either steam
or hot liquid water (HLW) see Table 13 which is the likely solvent of choice for biomass
conversion because it is cheap and readily dissolves significant amounts of polar oxygenate
compounds from biomass (Ravenelle et al 2010)
The hydrothermal stability of zeolites in steam has been intensively investigated (Xiong
et al 2014) In the presence of steam zeolites may undergo dealumination in which the
framework aluminium atoms (FAl) are removed from the zeolite lattice (Sanz et al 1988
Triantafillidis et al 2000 Triantafillidis et al 2001 Van Donk et al 2003) The
dealumination may lead to a structural breakdown depending on the degree of dealumination
while the degree of dealumination depends on the type and SiAl ratio of zeolites as well as the
dealumination methods (Sanz et al 1988 Triantafillidis et al 2000 Triantafillidis et al
2001 Van Donk et al 2003) The Al atoms removed from the framework form extra-
framework aluminium (EFAl) species (Klinowski 1984) which may either reside separately
outside of the zeolite crystal (Sanz et al 1988 Stockenhuber and Lercher 1995 Ennaert et
al 2014) or be present as isolated charge-compensating cations in the zeolite cages see Figure
19 (Stockenhuber and Lercher 1995 Beyerlein et al 1997) The presence of EFAl in zeolites
results in several desired properties such as increased hydrothermal stability and Lewis acidity
(Ennaert et al 2014)
68
As opposed to the hydrothermal stability of zeolites in steam only a handful of studies
focused on the behaviour of zeolites in HLW at gt100 degC ie water in a condensed state The
studies on the hydrothermal stability of zeolites in HLW summarised in Table 13 are
discussed in this section
69
Table 13 A summary of studies on the hydrothermal stability of zeolites in hot liquid water (HLW)
Zeolite (zeolite type) - SiAl Temperature
Pressure Time
Note Ref
Zeolite Beta (BEA)-18 38 360
Mordenite (MOR)-20
Zeolite Y (FAU)-30
85 ˚C na 5 h Solubility of the zeolites decreased with increasing SiAl ratio
H-BEA-360 lt H-BEA-38 lt H-BEA-18
The solubility was affected by the framework type (i) H-MOR-
20 lt H-BEA-18 (ii) H-BEA-30 lt H-FAU-Y-30
Kruger et al (2014)
Zeolite Y (FAU)-5 14 41
ZSM-5 (MFI)-15 25 40
150 ˚C 200 ˚C
saturated vapour
pressure 1-6 h
All ZSM-5 were stable whereas zeolite Y samples were
degraded and transformed into amorphous
Ravenelle et al (2010)
Zeolite Y (FAU)-27 130 to 200 ˚C
saturated vapour
pressure 72 h
waterzeolite ratio =
100
Decomposition of zeolite H-Y increased with increasing
temperature The zeolite slowly decomposed into EFAl and Si
species at ge 150 ˚C while at 170 and 200 ˚C its structure rapidly
collapsed transforming into Kaolinite amorphous SiO2 and
metakaolinite
Dimitrijevic et al (2006)
ZSM-5 (MFI)-24 to 918
mordenite (MOR)-12 20
Zeolite Beta (BEA)-21 to 200
Zeolite Y (FAU)-5 to 265
130-240 ˚C saturated
vapour pressure 72
h waterzeolite ratio
= 50
ZSM-5 and mordenite were stable whereas zeolite Y and Beta
decomposed
Lutz et al (2005)
H-Y (FAU)-25
H-USY (FAU)-19 9 3
35 ppm of HCl 190
˚C under stirring at
750 rpm 50 bar H2
24 h
all US(Y) zeolites dissolved but showed structural stability for
24 h USY19 has the highest stability even after 120 h
Ennaert et al (2014)
ZSM-5 (MFI)
mordenite (MOR)
Zeolite Beta (BEA)
Zeolite Y (FAU)
200 ˚C saturated
vapour pressure vary
reaction time
The main factor determining the structural stability of zeolites
is the density of their Si-OH terminated defects
Zhang et al (2015)
70
Zeolites tend to be decomposed by water under hydrothermal conditions The degree
of decomposition of zeolites in HLW depends on several factors such as the type and the SiAl
ratio of zeolites see Table 13 Lutz et al (2005) investigated the structural stability of ZSM-5
(MFI) mordenite (MOR) zeolite Beta (BEA) and zeolite Y (FAU) of different SiAl ratios
The zeolites were exposed to the CO2-free hydrothermal environment at 130-240 ordmC under
saturated vapour pressure for 72 h at the waterzeolite ratio of 50 ZSM-5 and mordenite were
relatively stable under the conditions tested whereas zeolite Y and Beta decomposed The
higher stability of ZSM-5 and mordenite was attributed to their dense framework structures
The low stability of Zeolite Y (FAU) under hydrothermal conditions was in a good
agreement with the results reported by Dimitrijevic et al (2006) and Ravenelle et al (2010)
Dimitrijevic et al (2006) reported that the decomposition of zeolite H-Y (SiAl = 27) increased
with increasing temperature The destruction of the zeolite into EFAl and Si species was visible
at 150 degC while the structure collapse to kaolinitemetakaolin and amorphous substances SiO2
was strongly accelerated at the temperatures between 170 and 200 degC The transformation of
zeolite Y into amorphous solid was also observed by Ravenelle et al (2010)
The structural stability of Beta zeolite in hot liquid water is also low As reported by
Lutz et al (2005) Beta zeolite decomposed in liquid water at 130-240 degC saturated vapour
pressure 72 h Kruger et al (2013) analysed the reaction solution recovered from the aqueous
phase dehydration of fructose catalysed by H-Beta zeolite (SiO2Al2O3 ratio = 18) at 130 degC
for 5 h finding that 1-3 wt of the total zeolite dissolved into Al and Si species The zeolite
continued to generate aqueous Si and Al species even after the 2nd and 3rd use However it
should be noted that the researchers cooled down the mother liqueur to laboratory temperature
prior to sampling and analysis affecting a temperature dependent equilibrium concentration of
aqueous species due to precipitation Kruger et al (2014) subsequently studied the dehydration
of fructose carried out in aqueous solutions pre-equilibrated with various zeolites at 85 degC with
71
or without formic acid (or HCl) at pH = 3 The reason as to why the acid was added was to
mimic the conditions in which H-Beta zeolite present and the acidic products formed in the
reaction may influence the zeolite dissolution It was found that the solubility of the zeolites
decreased with increasing SiAl ratio H-BEA-18gt H-BEA-38gt H-BEA-360 where the
number indicates the SiO2Al2O3 ratio The solubility was also affected by the framework type
(i) H-BEA-18gt H-MOR-20 (ii) H-FAU-Y-30gt H-BEA-30
ZSM-5 (MFI) and mordenite (MOR) exhibit high stability in water at 130-240 degC (Lutz
et al 2005 Ravenelle et al 2010) but their long-term stability may be decreased when
subjected to such conditions repeatedly andor for longer time Another group of zeolites stable
in HLW is ldquoUltra stablerdquo Y zeolites (USY) Ennaert et al (2014) examined the hydrothermal
stability of H-Y (SiO2Al2O3 = 25) and H-USY (SiO2Al2O3 = 19 9 and 3) in water acidified
with 35 ppm of HCl at 190 degC under stirring at 750 rpm 50 bar H2 24 h Despite some degree
of dissolution all zeolites were stable under the conditions applied USY with SiO2Al2O3 of
19 exhibited the highest stability and maintained its structure even after 120 h The stability of
the (US)Y in HLW is influenced by the presence of FAl and EFAl species formed either by
steaming or treatment in HLW on the external surface The FAl protects zeolite lattice from
hydrolysis while the EFAl exhibiting low water solubility prevents solubilisation of zeolite
framework
It is well established that the dealumination of zeolites is enhanced by acids while the
desilication is facilitated by bases (Čimek et al 1997 Dimitrijevic et al 2006) Under
hydrothermal conditions the dissolution of zeolites may be facilitated by the presence of both
H+ and OH- ions Figure 20 and 21 show zeolite decomposition steps under hydrothermal
conditions The Si-O-Al and Si-O-Si bonds are attacked by the H+ and OH- ions of water
molecules respectively The latter usually starts at the terminal OH-groups on the surface
72
Figure 20 Decomposition of a low-silica zeolite framework by water Adapted from Lutz et
al (2005) and Ravenelle et al (2010)
Figure 21 Decomposition of a high-silica zeolite framework by water Adapted from Lutz et
al (2005)
73
The final products of zeolite decomposition include SiO2 Si(OH)4 and Al(OH)3 Other
products such as aluminosilicate species and amorphous solids andor crystalline SiO2 may
also be present (Čimek et al 1997) The solubility of SiO2 in water is generally low but
gradually increases with temperature and pressure up to the neighbourhood of the critical
temperature of water where the solubility changes significantly at relatively low pressures
with increasing temperature the solubility is very low (Brunner 2009) SiO2 dissolves in water
to H4SiO4(s) which is subsequently hydrolysed to form silicic acid H4SiO4(aq) see Equation 4
(Rimstidt and Barnes 1980)
Equation 4 Equilibrium of SiO2 in water
1198781198941198742(119904) + 21198672119874(119897) harr 11986741198781198941198744(119904)helliphelliphelliphelliphelliphelliphelliphellip (i)
11986741198781198941198744(119904) + 1198672119874(119897) harr 11986731198781198941198744minus(119886119902) + 1198673119874+helliphelliphelliphellip(ii)
Overall reaction 1198781198941198742(119904) + 21198672119874(119897) harr 11986741198781198941198744(119886119902)helliphelliphelliphelliphelliphelliphellip(iii)
The dissolution of zeolite Y in HLW has been proposed to proceed via hydrolysis of
Si-O-Si bonds (Ravenelle et al 2010) Increasing the SiAl ratio leads to an increase in its
hydrophobicity and thus increasing its stability in aqueous environments Zapata et al (2012)
reported that the structural stability of H-Y zeolites in HLW at 200 degC may be increased by
functionalization of the surface of Y-zeolites with hydrophobic octadecyltrichlorosilane This
increased the hydrophobicity of the zeolites without considerably reducing the density of acid
sites This approach was also adopted by Zhang et al (2015) who demonstrated that the main
factor affecting the stability of zeolites in HLW is as opposed to other studies the density of
74
Si-OH terminal groups The silylation of the Si-OH with ethyltrichlorosilane increased the
hydrophobicity and the stability in HLW
233 Zeolite catalysts in the hydrothermal conversion of polyols to
chemicals
Despite the generally low stability of zeolites in HLW (Ravenelle et al 2010 Kruger
et al 2014) several studies utilized zeolites catalysts for hydrothermal conversion of glycerol
(see Table 10 page 53) and other carbohydrate biomass materials (Neher et al 1995 Kato
and Sekine 2013 Dornath and Fan 2014 Gonzalez-Rivera et al 2014) Table 14 summarises
the research studies including experimental conditions and results on zeolite-catalysed
hydrothermal conversion of various carbohydrates into chemicals
75
Table 14 A summary of research studies including experimental conditions and results on hydrothermal conversion of various polyols into
chemicals using zeolite-based catalysts The percentage values are given in mol unless stated otherwise
Feedstock Temperature
Pressure Time
Catalysts Key products
Ref
Cellulose (150 170 190 ˚C)
na (1 3 5 h)
PtH-Beta CO CO2 C1-C5 hydrocarbon
furfural HMF levulinic acid
Kato and Sekine (2013)
Cellulose 210 ˚C 18 bar 30
min microwave-
assisted
H-Beta Sn- Zn- K-doped
Beta zeolite
glucose HMF levulinic acid acetic
acid lactic acid
Gonzalez-Rivera et al (2014)
Fructose (120 150 165 ˚C)
autogenous 05-2 h
H-Beta with carbon black as
an adsorbent
Y (HMF amp furfural) = ~40 Dornath and Fan (2014)
Fructose 130 ˚C na 5 h H-Beta zeolite Sn-Beta
zeolite H-Beta zeolite
filtrated
HMF levulinic acid glucose Van Donk et al (2003)
Fructose 130 ˚C na 5 h H-Beta zeolite zeolite
filtrate of Beta zeolite
Mordenite and zeolite Y
HMF levulinic acid glucose Kruger et al (2014)
DHA GLA 125 ˚C na 24 h Sn-Beta Zr-Beta Ti-Beta
Al-Beta Si-Beta
Y (lactic acid) = ~90 Taarning et al (2009)
Note S = selectivity Y = yield na = not available but assumed to be saturated vapour pressure HMF = 5-hydroxymethylfurfural DHA =
dihydroxyacetone GLA = glyceraldehyde LHSV = Liquid hourly space velocity
76
Dornath and Fan (2014) investigated the hydrothermal dehydration of fructose over H-
Beta zeolite at 120 150 165 ordmC autogenous pressure 05-2 h The yield of HMF the main
product depended mainly on its subsequent conversion to formic acid and levulinic acid see
Figure 22 To increase the yield of HMF a carbon black (BP2000) was used to selectively
adsorb the HMF and furfural to prevent them to react further thus increasing their selectivity
from 27 mol to 44 mol equivalent to the total yield of 41 mol The catalytic effect of
dissolved aqueous zeolites species was studied by Kruger et al (2013) (2014) who investigated
aqueous phase dehydration of fructose using H-Beta zeolite finding the zeolite to be acting as
a source of aqueous aluminosilicate species adversely affecting the selectivity to the desired
products 1 to 3 wt of the total zeolite used dissolved at water-to-zeolite mass ratio ~27 at
130 degC However it should be noted that the researchers cooled down the mother liqueur to
laboratory temperature prior to analysis possibly affecting the equilibrium concentration of
aqueous species due to precipitation The dehydration of fructose was also carried out in
solutions previously equilibrated at 85 degC with zeolites of various framework types and
SiO2Al2O3 ratios (Kruger et al 2014) The solubility of zeolites decreased with increasing the
SiAl ratio H-BEA-18gt H-BEA-38gt H-BEA-360 where the number indicates the SiO2Al2O3
ratio The solubility was also affected by the framework type (i) H-BEA-18gt H-MOR-20 (ii)
H-FAU-Y-30gt H-BEA-30
Figure 22 Dehydration of fructose to HMF (adapted from Dornath and Fan (2014))
77
The zeolites modified to exhibit increased Lewis acidity exhibited promising activity
in converting of biomass into chemicals For instance Kato and Sekine (2013) studied the
hydrothermal conversion of cellulose at 150 170 and 190 ordmC for 1 3 and 5 h in the presence
of M-doped γ-Al2O3 and M-doped H-zeolite with various SiO2Al2O3 ratios where (i) M = Pt
Pd Co Cu Ni or Fe and (ii) H-zeolite = H-Beta zeolite H-USY H-MOR H-ZSM-5 or H-
FER The metal-doped catalysts were prepared by impregnation and liquid ion exchange (LIE)
with various target metal loadings In the presence of γ-Al2O3 and H-zeolites the liquid
products detected were furfural HMF and levulinic acid On the other hand the use of metal-
doped catalysts also provided gases including C1 to C5+ hydrocarbons The 19 wt Pt
supported on H-Beta zeolite (SiO2Al2O3 = 25) prepared by LIE provided the highest
conversion of cellulose and selectivity to valuable C3 and C4 hydrocarbons
Gonzalez-Rivera et al (2014) reported that the microwave-assisted hydrothermolysis
(210 degC autogenous pressure) of microcrystalline cellulose catalysed by H-Beta zeolite
resulted in several organic compounds such as glucose 5-HMF levulinic acid acetic acid and
lactic acid Zn-Beta zeolite favoured the glucose formation and performed better than Sn-Beta
and K-Beta zeolites however none of them achieved the performance of H-Beta zeolite
possibly due to lower concentration of Broslashnsted acid sites On the other hand Lewis-acidic Sn-
Beta zeolites were reported to have high activities and selectivities in the isomerization of
trioses to lactic acid 90 mol yield of lactic acid from hydrothermolysis of DHA at 125 degC
was obtained (Taarning et al 2009)
In spite of the promising catalytic activities and low prices of transition metals as
compared to precious metals no study reported the use of zeolites doped with these types of
Lewis acidic metals in the hydrothermal conversion of glycerol The present study focuses on
78
the use of transition metals-doped ZSM-5 and Beta zeolites as catalysts for the hydrothermal
transformations of glycerol Both zeolites were selected for their high Broslashnsted acidity and ion
exchange capacity (Weitkamp 2000) The metals selected for doping of the zeolites in this
work included Sn2+ and Zn2+ due to their high catalytic properties in the hydrothermal
transformations of polyols (Taarning et al 2009 Cho et al 2014 Gonzalez-Rivera et al
2014) Other investigated metals were Cu2+ Ce3+ and La3+ The Cu2+ Ce3+ and La3+-doped
zeolites have never been reported in the hydrothermal conversion of polyols but these high
Lewis acidic metals were used in other forms showing promising activity For instance Cu2+
was used for the hydrothermolysis of glycerol in sub-CW as CuOAl2O3 at high pH providing
a 98 mol conversion with a high selectivity to lactic acid (~79) at 240 degC after 6 h (Roy et
al 2011) Similarly the hydrothermal conversion of cellulose at 250 degC catalysed by 30 micromol
LaCl3 provided 80 conversion in 180 s with HMF as the main product (Seri et al 2002)
Ce3+ was selected as another representative of lanthanides because among all the lanthanides
it is most abundant and reactive
234 Preparation of Lewis acidic zeolite catalysts
The modification of zeolites through the ion exchange reactions is often carried out in
liquid phase with a high yield of ion exchanged zeolite However the method may have several
disadvantages (Kinger et al 2000)
i The steric constraints may exist due to the formation of bulky hydration shells of
the cations to be exchanged meaning that the diffusion of the cations into the zeolite
pores is limited
ii The solubility of the cation precursor may be low limiting the amount of cations
available in solution
79
iii The extent of exchange is limited by the thermodynamic equilibrium The exchange
process may need to be repeated several times to achieve high ion exchange degree
resulting in the need for large volumes of solution generating waste
The above limitations may be avoided by using an alternative method known as solid-
state ion exchange (SSIE) involving (i) mechanical grinding of powder zeolite with a salt of
desired cations and (ii) calcining the mixture at high temperature SSIE of zeolite is an efficient
process and its main advantage over liquid ion exchange is that a high degree of exchange may
be achieved in a single-step treatment (Kinger et al 2000) Also the metal containing zeolites
prepared by SSIE and conventional liquid ion exchange showed comparable activity and
selectivity in a number of reactions such as NiH-MFI and NiH-BEA in the n-nonane
hydroconversion (Kinger et al 2000) Ce-ZSM-5 in the selective catalytic reduction of NOx
(van Kooten et al 1999)
The metal-doped zeolites prepared by the ion exchange of the balancing cation exhibit
weak or moderate Lewis acidity because many of the exchangeable cations have a low
electronegativity (eg alkali or alkali earth elements) or have a relatively large atomic diameter
eg La3+ Also the balancing metals may be leached out of the frameworks These issues may
be alleviated by incorporating a desired metal into the frameworks replacing the aluminium
atoms Metal-incorporated zeolites are typically obtained by direct incorporation of the desired
metals into the frameworks during hydrothermal synthesis This process however has several
drawbacks (Hammond et al 2012) such as
i) the formation of metal oxide particles which are significantly less active than the
metal ions in the zeolite framework
ii) long synthesis times with unfavourably large crystals formation and
iii) the need for dangerous additives such as hydrogen fluoride
80
Several research groups have developed a convenient post-synthetic method for the
incorporation of different Lewis acid metals into the framework without changing the
crystalline structure of zeolite (Corma et al 1998 Hammond et al 2012) The method
involves the dealumination of zeolites by strong inorganic acid without collapsing the zeolite
structures and filling of vacated tetrahedral sites with metal atoms by a SSIE reaction This is
done by contacting the dealuminated zeolite with a desired metal salt at high temperature (ge500
degC) The method not only eliminates the long synthesis times required in the conventional
hydrothermal synthesis routes but also yields smaller crystals than previously possible through
the direct synthesis The metal-doped zeolites were demonstrated to have a high catalytic
performance in various reactions Cu-Beta zeolites in the selective catalytic reduction (SCR)
of NOx (Corma et al 1997) Sn-Beta zeolite Sn-MCM-41 Ga-MCM-41 catalysed
isomerisation of triose sugars (Taarning et al 2009 Li et al 2011) Sn-deAl-Beta zeolite in
the BaeyerndashVilliger oxidation of cyclohexanone (Hammond et al 2012) However to the best
of our knowledge they have not been exploited for the hydrothermal conversion of glycerol
235 Immobilisation and shaping of powder zeolite catalysts
Powdered zeolites are unsuitable for a direct use in a continuous flow reactor due to
their nano-sized particles having a tendency to get washed out It is therefore necessary to
immobilise them on a larger particle support or agglomerate the powdered zeolites into
macroscopic forms such as pellets or extrudates prior to administration
2351 Immobilisation of zeolites on inorganic membrane (zeolite membrane)
Immobilisation of zeolites on an inorganic membrane consisting of a zeolite top layer
on a mesoporous ceramic or metal support has been developed over the last two decades (Caro
81
et al 2000 Caro and Noack 2008 Gascon et al 2012) as an alternative to polymer composite
membranes widely used in separation operations but their applications are limited due to their
poor thermal and chemical resistance poor durability and catalytic deactivation (Shelekhin et
al 1992 Ulbricht 2006 Yampolskii 2012) In contrast zeolite membranes are relatively
robust offering the possibility of continuous operation without recurring regeneration and thus
reducing the process downtimes NaA zeolite membrane has been used in a plant with 530
Lmiddoth-1 production capacity of solvents at less than 02 wt of water from the solvents initially
contained 10 wt H2O at 120 ordmC (Morigami et al 2001) The use of zeolite membranes in
catalytic reactions was investigated by (McLeary et al 2006 Van Den Bergh et al 2010)
The zeolite membranes may be viewed as a single unit molecular reactors acting as
both catalysts and separators (McLeary et al 2006) High conversion and selectivity were
reported for several types of reactions Hasegawa et al (2001) reported nearly complete
oxidation of CO in H2-rich mixtures using Pt-loaded Y-type zeolite membrane supported on
porous α-Al2O3 tube at 200-250 ordmC demonstrating the potential application of zeolite
membranes for fuel cells Barbieri et al (2002) investigated the potential of zeolite membrane
reactors as alternatives to a fixed bed tubular reactor for the hydrogenation of CO2 to CH3OH
Both the organophilic membrane (such as a silicalite-1 membrane) and hydrophilic membrane
(such as the MORZSM-5chabazite membrane) can operate at lower pressure and higher
temperature as compared to a fixed bed tubular reactor providing higher conversion (40-50)
and selectivity (almost 100) The higher performance of the membrane reactors possibly
because of the selective product removal shifts the reaction equilibrium towards the products
Despite high performance zeolite membrane reactors possess low packing density
(130-800 m2m3) (ie membrane separation areamodule volume ratio) (Xu et al 2004) To
overcome this issue the immobilisation of zeolites on ceramic hollow fibre membranes has
been investigated Ceramic hollow fibre membranes made of Al2O3 have a micro-tubular
82
configuration (outside diameter lt2 mm) and very high packing density (1000-9000 m2m3) (Xu
et al 2004 Tan and Li 2011) see Figure 24
Figure 24 SEM micrographs of ceramic hollow fibre membranes (taken from Gbenedio et al
(2010))
The high temperature resistance as well as high chemical and mechanical stability make
the ceramic hollow membranes an attractive choice for the reactions andor processes operating
in harsh environments (Kingsbury and Li 2009) such as filtration of corrosive fluids (Weber
et al 2003) water desalination (Fang et al 2012) solid oxide fuel cells (SOFCs) (Wei and
Li 2008) membrane contactors (Koonaphapdeelert and Li 2006) membrane supports (Julbe
et al 2001) and high temperature membrane reactors (Galuszka et al 1998 Neomagus et al
2000 Keuler and Lorenzen 2002 Gbenedio et al 2010 Garciacutea-Garciacutea et al 2012)
Immobilised zeolites on ceramic hollow membranes were successfully prepared and
used in several gas and liquid separations Alshebani et al (2008) grown the MFI zeolite into
the ceramic hollow fibres displaying the separation factors close to 10 for the CO2H2 mix The
mixes tested by other researchers include for instance a mix of n-butaneH2 (Li et al 2008)
83
and xylene isomers (Daramola et al 2009) Xu et al (2004) deposited a considerable amount
of NaA zeolite on a ceramic hollow fibre which could be used in gas separations However to
the best of our knowledge zeolites on ceramic hollow membranes has not been reported for
liquid phase reactions
2352 Shaping of zeolite powder
A majority of industrial heterogeneous catalysts are usually employed in their
macroscopic forms such as pellets granules and extrudates (Kraushaar-Czarnetzki and Peter
Muumlller 2009) The most common industrial shaping methods include (Mitchell et al 2013)
spray drying - used to prepare micrometre-sized agglomerates
pelletisation extrusion or granulation - used to prepare millimetre- or centimetre -sized
agglomerates
Depending on the method used a mixture of powdered zeolite and additives such as binders
peptisers and plasticiser may be processed differently (i) dry mixes may be pelletised while
(ii) wet mixes may be extruded granulated or spray dried (Mitchell et al 2013)
Although it may be possible to press powdered zeolites into pellets without any additive
(Mikkola et al 2014) the cohesion of zeolites alone cannot provide sufficient strength for the
catalysts to be used in industrial practice Binders are the materials usually added to improve
mechanical stability and attrition resistance of the catalysts upon hardening (usually achieved
by thermal treatment) of the shaped bodies (Sntamaaria et al 2002 Bauer et al 2014) Silica
alumina and natural clays such as bentonite are among the most commonly used binders for
zeolites see Table 15 Other additives may also be used for instance a peptiser for dispersing
particles to improve the feed homogeneity prior toduring shaping and plasticiser for decreasing
the viscosity of pre-mixtures to facilitate the processing (Mitchell et al 2013)
84
Table 15 Examples of zeolite-based catalysts agglomerated with commonly used binders and additives The reactions catalysed by the catalysts
prepared are also presented
Zeolite Binder Other additives Tested reactionreaction conditions Ref
Honeycomb H-ZSM-5 colloidal silica methyl cellulose
(plasticiser)
oxidation of 12-dichloroethane
at 200-450 degC gas phase flow reactor
Aranzabal et al (2010)
Pt-Pd supported on Beta zeolite
(SiAl = 125 or 25)
γ-Al2O3 HNO3
(peptiser)
hydroisomerization of long-chain n-paraffins
at 205-230 degC pH2 = 50 bar 1198991198672119899ℎ119910119889119903119900119888119886119903119887119900119899 =
7501 liquid hourly space velocity (LHSV) of 05 hminus1
Bauer et al (2014)
H-Beta zeolite Al2O3 acetic acid
(peptiser)
isopropylation of benzene
at 21 degC WHSV 25 h-1 feed (benzeneisopropyl
alcohol) 651
Kasture et al (2007)
H-Gallosilicate (MFI) zeolite Al2O3 kaolin - propane aromatization
at 550 degC gas phase continuous flow reactor time-on-
stream about 6 h
Choudhary et al (1997)
Ti-SBA-15 bentonite methyl cellulose
(plasticiser)
liquid phase olefin epoxidation
at 110 degC 2 h in batch at 110120 degC in a flow reactor
Melero et al (2008)
85
The formulation and the particle shape may affect the performance and mechanical
strength of the agglomerated zeolites Kasture et al (2007) studied the isopropylation of
benzene using Beta zeolite with different amounts of alumina binder The Beta zeolite was
prepared from the fly ash while the binder was added to the acidic forms of the prepared zeolite
at the dry weight ratio of sample to binder of 9010 8020 7030 6040 and 5050 The final
composite material was pressed into self-bonded pellets as well as extruded Among all the
catalysts prepared the 40 wt alumina binder was found to be the most active catalyst with
respect to the benzene conversion and the selectivity to isopropylbenzene an intermediate for
the production of phenol ~19 conversion with ~94 selectivity was obtained using the
pelletized catalyst while the reactions with the extrudates provided ~18 conversion with 92
selectivity
Melero et al (2008) extruded Ti-SBA-15 with bentonite clay and examined its
performance on the liquid phase olefin epoxidation The bentonite content of the extrudates
varied from 10-40 wt with 30 wt being optimal with respect to the particle resistance and
the Ti content Larger amounts of bentonite did not increase the particle mechanical resistance
but decreased the activity of the final material due to the dilution of the Ti-SBA-15 catalyst
A binder typically accounts for 50-90 wt of agglomerated zeolite catalysts However
certain binders may chemically interact with zeolites and act as catalysts (Choudhary et al
1997) As a consequence the overall activity selectivity and stability of the zeolite may be
influenced by the binder-zeolite interactions (Choudhary et al 1997 Zhang et al 2006 Zhao
et al 2013) For instance Choudhary et al (1997) investigated the influence of alumina and
kaolin on the acidity and performance of a commercial H-Gallosilicate (MFI) zeolite (H-
GaMFI zeolite) on the propane aromatization The zeolite was mechanically mixed with a
binder pressed and crushed into ~02-03 mm particles prior to the calcination at 600 degC The
total concentration of strong acid sites of the agglomerated H-GaMFI zeolite determined by
86
chemisorption of pyridine at 400 degC was reduced significantly when kaolin binder was used
while alumina binder did not significantly affect the total acidity The content of binder tested
at 10-50 wt had a negligible effect on the concentration of the acid sites The conversion of
propane and the selectivity to aromatics in the reactions catalysed by the catalyst agglomerated
with kaolin were decreased whereas no change was observed when alumina binder was used
2353 Agglomeration of zeolite powder
The low intracrystalline diffusion coefficient of zeolites (ranging from 10-8 to10-18 m2s)
(Landau et al 1994) usually results in coke formation and rapid catalyst deactivation making
the process optimization difficult (Bonardet et al 1999 Argyle and Bartholomew 2015) A
possible solution to overcome this mass transport limitation is to decrease the zeolite crystallite
size to nanoscale leading to higher surface area to micropore volume ratio and shorter diffusion
path length (Yamamura et al 1994) However the direct use of nano-sized zeolites may cause
high pressure drops in packed-bed reactors up-flow clogging of equipment as well as health
and environmental issues (Waller et al 2004) Therefore nano-sized zeolites need to be
agglomerated or dispersed on a porous support
Typically zeolites are mixed with a binder and subsequently extruded or compressed
into beads or pellets (Kraushaar-Czarnetzki and Peter Muumlller 2009) However binders are
usually not designed and structured for optimal mass transport in terms of high specific surface
areas and pore geometry (Choudhary et al 1997 Ramos et al 2008 Hargreaves and
Munnoch 2013) Generally zeolites agglomerated with a binder possess macro- meso- and
micropores with total BET surface areas of only around 300 m2middotg-1 (Waller et al 2004)
An alternative approach investigated by a number of researchers is to synthesize
zeolites inside the pores of mesoporous supports (Landau et al 1994 Madsen et al 1999)
87
Embedding of zeolite nanocrystals into a mesoporous inorganic matrix such as MCM-
n and TUD-1 has gained increasing interest as a method for agglomerating zeolites (Jansen et
al 2001 Waller et al 2004 Xia and Mokaya 2004 Mavrodinova et al 2005 Petkov et al
2005 Wang et al 2007 Wang et al 2008 Wang et al 2009 Xu et al 2009 Lima et al
2010) The high surface areas and uniform mesoporosity of the support allow the composites
to benefit from the high zeolite catalytic activity and high diffusion rate Wang et al (2007)
(2008) (2009) developed a one-pot synthesis method to obtain a ZSM-5 zeolitemeso-matrix
composite with tunable mesoporosity and high BET surface area (~600 m2g) Xu et al (2009)
synthesized BetaMCM-41 composite exhibiting high BET surface area (~830 m2g) and high
hydrothermal stability its mesoporous structure was preserved even after 336 h in boiling
water Waller et al (2004) prepared Beta zeolite nanocrystals inside mesoporous TUD-1
exhibiting a higher activity per gram of zeolite than pure nanosized Beta zeolite for n-hexane
cracking The nature of the acid sites in the nanozeolite was partially modified due to the
interactions with the TUD-1 Lima et al (2010) selected TUD-1 as a matrix to agglomerate
with H-Beta zeolite The Beta zeoliteTUD-1 composite was used in the cyclodehydration of
xylose to furfural in a H2O-toluene mixture achieving at 170 ordmC the xylose conversions of
~98 higher furfural yield (74) as compared to that obtained on bulk H-Beta zeolite (54)
Several other zeoliteTUD-1 composites were also reported for instance zeolite
YTUPD-1 composite prepared for benzylation reactions (Hamdy and Mul 2013) and ZSM-
5TUD-1 composite used in aldol condensation (Zhou et al 2010)
88
Table 16 Applications of zeoliteTUD-1 composites reported in literature
Reaction Conditions Composite Results Reference
Cracking of n-
hexane
538 ordmC Beta
zeoliteTUD-1
k = 025 while 011
for pure Beta zeolite
Waller et al
(2004)
aldol condensation
of benzaldehyde
and n-amyl
alcohol
80 ordmC ZSM-5TUD-1 X = 31 while 12
for our ZSM-5
Zhou et al
(2010)
cyclodehydration
of xylose
H2O-toluene
mixture 170
ordmC
H-BetaTUD-1 X = 98 furfural
yield = 74 while X
= 94 and furfural
yield = 58 for a
mixture of Beta and
TUD-1
Lima et al
(2010)
Friedel-Crafts
benzylation
80 ordmC H-YTUPD-1 k = 24 while 18 for
pure H-Y zeolite
Hamdy and
Mul (2013)
Pseudo-first-order reaction rate constant in gzeolite-1middotmin-1middotL first-order rate constant in h-1middotmg-1 X
conversion
89
3 Aim of study
31 Aim
The surplus of crude glycerol from the biodiesel industry has resulted in a significant
drop in its price making it an interesting feedstock for the production of value-added chemicals
Finding new uses of crude glycerol would lead to (i) a more efficient biodiesel production
process by reusing its current waste product and ultimately (ii) to a more competitive biodiesel
fuel economy The aim of this thesis is to improve our understanding of the hydrothermal
conversion of glycerol facilitated by the zeolite-based catalysts with respect to
(i) their activity in the hydrothermal transformations of glycerol and
(ii) their potential to affect the distribution of reaction products in favour of liquid
conversion products of glycerol
The hydrothermal environment was adopted for its ability to provide rapid reaction
rates and the high water content of crude glycerol (Peterson et al 2008) ZSM-5 and Beta
zeolites were selected for their high Broslashnsted acidity and ion exchange capacity (Weitkamp
2000) The metal ions selected for doping of the zeolites included Ce3+ Cu2+ La3+ Sn2+ and
Zn2+ due to their promising activity in the hydrothermal transformations of polyols see Chapter
233
In order to achieve the aim of the project the following five objectives are specifically
addressed
90
32 Objectives
Objective I The effect of key reaction parameters on selectivity and yield
This study initially focuses on the effect of high temperaturepressure water on the
decomposition of glycerol in a binary system glycerol-water Also the presence of zeolite with
and without the presence of H2O2 as an oxidation agent is taken into account The individual
systems studied in the initial phase (batch mode) of current work are as follows
(i) Glycerol - water
(ii) Glycerol - water - non-modified commercial catalyst (H-Beta zeolite H-ZSM-5)
(iii) Glycerol - water - H2O2
(iv) Glycerol - water - H2O2 - non-modified commercial catalyst
The ranges of key parameters used
(i) Reaction time 1-300 min
(ii) Pressure 23-186 bar
(iii) Temperature 105-360 degC
The selection of suitable reaction pressure temperature and time is essential since for
instance a short reaction time may lead to a low conversion while an excessively long reaction
time may lead to a high conversion lowering the yield of liquid products due to degradation of
primary products to gases The reaction time and temperature will be varied and the
composition of reaction products will be determined by quantitative and qualitative chemical
analysis 1H-NMR 13C-NMR 2D-NMR HPLC GC-FID and MS (for gaseous products) The
data collected will allow the determination of optimum reaction temperature and time leading
91
to a maximum conversion to liquid decomposition products of glycerol Such reaction
conditions will be used as a base line in further trials
Objective II Preparation and characterisation of metal-doped zeolites
The composition of glycerol conversion products not only depends on the temperature
pressure and reaction time but also on the type of catalyst(s) used This work focusses on the
use of heterogeneous zeolite catalysts exhibiting high Broslashnsted acidity and adjustable
BroslashnstedLewis acidity ratio (Almutairi 2013) The Broslashnsted acid sites of two selected
zeolites ie H-ZMS-5 and H-Beta zeolite will be transformed into Lewis acid sites by
replacing the balancing H+ with Ce3+ La3+ Sn2+ Zn2+ and Cu2+ using the solid state ion-
exchange technique (SSIE) In the case of H-Beta zeolite the transformation of Broslashnsted acid
sites into Lewis acid sites will also be carried out by replacing the Al atoms inside the
framework with Ce3+ La3+ Sn2+ Zn2+ See Table 22 23 and 24 for the zeolites and modified
zeolites prepared
The prepared catalysts will be characterised by several analytical techniques X-ray
powder diffraction spectroscopy (XRD) will be used to qualitatively identify crystalline
mineral phases in the samples while the content of Si and Al will be quantified by X-ray
fluorescence spectroscopy (XRF) The transmission and scanning electron microscopies (TEM
and SEM respectively) will be employed to image the catalystsrsquo morphology as well as to
study their elemental composition using energy dispersive X-ray spectroscopy analysis (EDS)
The catalysts surface area and pore volume will be determined by BrunauerndashEmmettndashTeller
analysis (BET) while the concentration of acidic active sites will be analysed by temperature
programed desorption of NH3 (TPD-NH3)
92
Objective III Investigation into the catalytic effect of zeolites and metal-
doped zeolites
The specific issues to be addressed in this objective are
The effect of metal-substitution of zeolites on hydrothermal transformations of glycerol
The potential of metal-doped zeolite catalysts prepared in this work to control the
distribution of conversion products of glycerol in favour of liquid products
An investigation into the catalytic effect of H-Beta zeolite H-ZSM-5 and metal-doped
zeolite catalysts will be carried out by varying the following reaction parameters temperature
reaction time and type of catalysts The performance of the catalysts will be evaluated by the
glycerol conversion rate selectivity to the major product as well as the stability of the catalysts
Due to the finely powdered nature of zeolite-based catalysts the initial experiments will
be performed in a batch mode with the 6 ml volume tubular bomb reactors as it is a rapid and
convenient method providing the desired information while eliminating the issues associated
with the use of continuous flow reactors such as the catalyst pellets disintegration and the
instrument maintenance
H2O2 will be used as a source of oxygen when required since it is an easy to use and a
powerful oxidant decomposing to oxygen and water eliminating the possibility of
contamination of the systems studied by foreign ions
Objective IV Design and modification of a continuous flow system
The selected reactions will be repeated using a continuous flow system equipped with
a packed bed reactor (PBR) also called plug flow reactor (PFR) In this type of reactors all the
93
substrates present at any given tube cross-section have identical residence time and so equal
opportunity for reaction Additionally in the continuous flow system the reaction parameters
such as temperature and pressure can be accurately measured and controlled All these features
allow the reactions to be performed more effectively at shorter residence times as compared to
a non-stirred batch process making it a more effective operating process for scaling up and
increasing the production capacity The finely powdered catalysts which are not suitable to be
used directly in the continuous flow system will be immobilised on a support pelletized or
extruded prior to administration
A bench-top continuous flow system will be designed and modified for the
hydrothermal transformations of glycerol
Objective V Kinetic study of glycerol transformation
A kinetic study on glycerol conversion under the conditions optimum for high
conversion degree and selectivity to liquid products will be carried out The reaction kinetics
reaction order reaction rate constant and the reaction activation energies will be determined
The data collected will allow the optimisation of the reaction conditions as well as improve the
understanding of the system
94
4 Experimental protocols and methodology
This chapter describes the experimental procedures used in this work together with
some theoretical background
41 Characterisation of solid samples
411 BrunauerndashEmmettndashTeller (BET) analysis
BET analysis is an instrumental technique used to determine the specific surface area
of solid materials by measuring of physical adsorption of a gas on the surface of a solid
substrate as a function of equilibrium pressure This technique is commonly used to determine
the specific surface area of zeolites and other microporous solid catalysts or catalyst supports
such as alumina and clay (Wenman 1995)
Theoretical background
The experimentally obtained adsorption isotherms ie the variation of adsorption with
pressure at a constant temperature are mathematically treated using a theory developed by
Stephen Brunauer Paul Emmett and Edward Teller (Brunauer et al 1938) The so-called
ldquoBET theoryrdquo is the most widely used theory to describe the phenomenon of multilayer
adsorption as opposed to Langmuirrsquos theory (Langmuir 1916) who introduced one of the first
physically plausible models explaining the adsorption The limitation of Langmuirrsquos model
constrain its use to the cases where there are negligible intermolecular interactions between
adsorbed particles ie a monolayer adsorption
95
In BET analysis the adsorption isotherms are treated using the Equation 5 The specific
surface area is found indirectly from the size and the volume of gas molecules needed to form
a monolayer on the surface of the sample (119881119898)
Equation 5 BET adsorption isotherm equation
1
[119881119886 (1198750
119875 minus 1 )]=
119862 minus 1
119881119898119862 times
119875
1198750+
1
119881119898119862
Where
119875 partial vapour pressure of adsorbate gas in equilibrium with the surface at -1958
degC (the boiling point of liquid nitrogen) Pa
1198750 saturated pressure of adsorbate gas at the temperature of adsorption Pa
119881119886 volume of gas adsorbed at standard temperature and pressure (STP) [0 degC and
atmospheric pressure of 1013 x 105 Pa] ml
119881119898 volume of gas adsorbed at STP to produce an apparent monolayer on the sample
surface ml
119862 dimensionless constant that is related to the enthalpy of adsorption of the
adsorbate gas on the powder sample
By plotting the BET value 1
[119881119886(1198750119875
minus 1)] against
119875
1198750 a straight line (usually in the
approximate relative pressure range from 005 to 03) is obtained From the resulting linear plot
the slope ( 119862minus1
119881119898119862 ) and the intercept (
1
119881119898119862 ) are obtained from the linear regression analysis Thus
96
the BET constant (C) Vm and the BET specific surface area (SBET) are calculated according to
Equation 6 7 and 8 respectively
Equation 6 The formula for the BET constant C
119862 = [119878119897119900119901119890
119868119899119905119890119903119888119890119901119905] + 1
Equation 7 The formula for the monolayer adsorbed gas quantity Vm (ml)
119881119898 = 1
119862 119868119899119905119890119903119888119890119901119905 119874119903
1
(119878119897119900119901119890 + 119868119899119905119890119903119888119890119901119905)
Equation 8 The formula for BET specific surface area SBET (m2gndash1)
119878119861119864119879 = (119881119898 119873 119886)
119898 22400
Where
N Avogadro number (6022 times 1023 molminus1)
Vm monolayer adsorbed gas quantity (ml)
a effective cross-sectional area of one adsorbate molecule m2 (0162 nm2
for N2)
m mass of adsorbent or solid sample g
22400 molar volume of adsorbate gas at STP ml
97
Experimental procedures
In present work the BET adsorption isotherms of samples were determined by using
nitrogen gas as a molecular probe at ndash1958degC and relative pressure (PP0) of 00 to 10 using
a TriStar 3000 (Micromeritics) analyser In a typical measurement approx 100 mg sample was
degassed in a flowing nitrogen gas at 200 degC for 4 h using a Flow prep 060 instrument
(Micromeritics) followed by evacuating and cooling to -1958 degC by liquid nitrogen Nitrogen
gas was dosed to the solid adsorbent in controlled increments After each dose of the gas the
pressure was allowed to equilibrate and measured The quantity of adsorbate ie the nitrogen
gas adsorbed on the adsorbent ie the solid sample was determined and plotted against the
equilibrium gas pressure The data were used to determine the quantity of gas required to
saturate the sample surface with a monolayer and the BET specific surface area (SBET) was
calculated using the quantity of the gas The micropore surface area and micropore volume
were calculated by the t-plot method
412 X-ray Fluorescence (XRF) spectroscopy
X-ray fluorescence spectroscopy is an analytical technique used for quantitative
elemental analysis of materials using an X-ray induced excitation of electrons
Theoretical background
When the X-rays collide with atoms or molecules their energy may be absorbed by core
electron resulting in either (i) its excitation to higher energy levels or (ii) ejection from the atom
leaving the atom in an excited state with a vacancy in the inner shell Outer shell electrons then
fill the vacancy emitting waves with energy equal to the energy difference between the two
energy states Since each element has a unique set of energy levels each element emits a pattern
98
of X-rays characteristic of the element termed ldquocharacteristic X-raysrdquo (Atkins and Paula
2006) see Figure 23 The intensity of the X-rays increases with the concentration of the
corresponding element The phenomenon is widely used for elemental analysis of solid
materials
Figure 23 The processes that contribute to the generation of X-rays An incoming electron
collides with an electron (in the K shell) and ejects it Another electron (from the L shell in
this illustration) falls into the vacancy and emits its excess energy as an X-ray photon The
figure was taken from Atkins and Paula (2006)
Experimental procedures
XRF analysis was carried out in a Bruker XRF Explorer-S4 analyser All samples were
subjected to the heating at 110 degC for 5 h in air atmosphere followed by 1000 degC for 5 h prior
to the X-ray fluorescence analysis The weight loss was recorded for both temperatures The
weight loss at 110 degC indicates the moisture content while the weight loss at 1000 degC also
99
referred to as the Loss On Ignition (LOI) may result from the presence of -OH groups andor
other volatiles eg S C and N compounds The calcined powder sample was placed into a
sample holder and analysed without further treatment
413 Powder X-ray diffraction (XRD) spectroscopy
Powder X-ray diffraction spectroscopy has been a principal instrumental technique
used for characterization of crystalline zeolite materials This non-destructive analytical
technique has invariably been the first choice for the identification of zeolites since it is rapid
and provides information on unit cell dimensions (Fultz and Howe 2013)
Theoretical background
X-ray diffraction analysis is based on the theory that crystalline materials with their
repeating structures can diffract radiation with a wavelength similar to the distances of lattice
planes in their structure (~1Aring) When the crystalline material is exposed to monochromatic x-
rays the incident beam is diffracted at specific angles to the lattice planes and can be detected
as a diffraction pattern characteristic of the material X-ray diffraction is at its best when dealing
with crystalline materials ie those with long range order
Braggrsquos law governs the conditions for diffraction (Atkins and Paula 2006) see
Equation 9
Equation 9 Braggs law
119899120582 = 2119889 sin 120579
100
where n is an integer
λ is the wavelength of the incident X-ray radiation
d is the distance between atomic layers in a crystal (d-spacing) and
120579 is the angle of incidence between the X-ray beam and the diffraction planes
Experimental procedure
X-ray powder diffraction (XRD) analysis was used for qualitative identification of
crystalline mineral phases in the samples The XRD spectra of samples were obtained on a
XPert PRO Diffractometer (PANalytical) using CuKα radiation (α = 15404 Adeg) and an
XCelerator detector with Nickel filter and Soller slit system All measurements were carried
out at laboratory temperature (22 degC) NaCl was used as an internal standard if required The
mass ratio of zeolite to NaCl was kept constant at 267 for all samples Unless otherwise stated
the angular scan was between 2θ = 5-75deg with a step size of 003deg 2θ and a count time of 38 s
per step
Phase identification was carried out using the PANalytical XPert High Score Plus
software The degree of crystallinity was determined based on the intensity of the characteristic
peak at around 2θ = 226deg (d302) for Beta zeolite and 2θ = 2303deg (d501) for ZSM-5 The
crystallite size of powder samples was determined by Scherrerrsquos equation see Equation 10
The instrument broadening of standard Si (111) at 2θ = 2844deg with CuKα1 radiation used was
015
101
Equation 10 Scherrer equation
119905 = 119870120582
119861 cos 120579
Where t is the particle size
K is the Scherrer constant typically ranging from 09 to 10 (~ 09 in this work)
λ is the x-ray wavelength (α = 154 Adeg)
Bstructural is the instrument broadening described by the full width at half
maximum (FWHM in radians) of the relevant peaks Bstructural = Bobserved -
Bstandard Bobserved is the observed peak width while Bstandard is the peak width of
a crystalline standard the instrument broadening of standard Si (111) at 2θ =
2844deg with CuKα1 radiation used was 015 (Demuynck et al 2010) FWHM
obtained from the machine is in degree it is therefore must be changed into
radian by multiplication by 120587180
120579 (theta) is half of the Bragg angle (in radians) The angle of diffraction obtained
is in degree and needs to be converted into radian by multiplication by 120587180
414 Electron microscopy
The microstructure of zeolites and supporting materials may be investigated effectively
by electron microscopy It provides information on the composition and morphology of
specimens
102
Theoretical background
Electron microscopy takes advantage of the interactions between a coherent beam of
electrons and a solid specimen The electron source is commonly a tungsten filament heated to
around 1700 degC at which point it emits copious electrons These electrons are accelerated across
a potential of several tens of keV to give them the required energy the beam is focused and
collimated by a sequence of electromagnetic lenses and apertures When the high energy beam
of electrons strikes the target specimens several interactions occur simultaneously The
incident electrons striking the target either (i) remain undeviated or (ii) are scattered and then
absorbed reflected (scanning electron microscopy) or transmitted (transmission electron
microscopy) Such interactions have led to the development of various electron microscopy
techniques including transmission secondary and backscattered electron microscopy all of
which are used to study the microstructure of materials Transmission electron microscopy and
Scanning electron microscopy were used to study the microstructure of catalysts in this work
Transmission electron microscopy (TEM)
Experimental procedure
A JEOL 2000-FX 200KV microscope equipped with an Energy-dispersive X-ray
spectrometer (TEM-EDX) was used A small amount of the powder samples were dispersed in
ethanol prior to depositing onto Holey carbon films on 300 mesh copper grids (Agar Scientific)
placed in sample holders
Energy-dispersive X-ray spectroscopy (EDX or EDS) is a spectroscopy technique used
in conjunction with SEM or TEM It detects the characteristic X-rays emitted by a sample
during bombardment with an electron beam An EDX spectrometer detector measures the
103
relative abundance of emitted x-rays versus their energy The spectrum of X-ray energy versus
counts was evaluated to determine the elemental composition of the samples
Scanning electron microscopy (SEM)
Experimental procedure
SEM was used to examine the crystallinity morphology and elemental composition of
powdered specimens using JEOL SEMEDX JSM-6010LA microscope The solid sample was
mounted in a silver sample holder on double-sided ldquoadhesive carbon taperdquo (Agar Scientific)
and gold coated
415 Temperature programed desorption of ammonia (TPD-NH3)
Temperature-programmed desorption (TPD) is an analytical technique used to measure
the gas adsorption capacity of a solid sample as a function of temperature TPD of basic
molecules such as NH3 (TPD-NH3) is one of the most commonly used methods for studying
the surface acidity of zeolites clays and mesoporous silica (Katada et al 1997 Niwa and
Katada 1997 Trombetta et al 2000 Flessner et al 2001 Rodrıguez-Castelloacuten et al 2003
Rodriacuteguez-Gonzaacutelez et al 2008) It was used to determine the number and strength of acid
sites of the catalysts tested in this work
TPD-NH3 is widely used for its simplicity and relatively low cost (Corma 1997)
However TPD-NH3 cannot distinguish between Broslashnsted and Lewis acid sites Such sites may
only be distinguished by Fourier Transform Infrared Spectroscopy coupled with pyridine
104
thermodesorption (FTIR-pyridine) However this may not suitable for samples with open
cavities or channels smaller than the molecules of pyridine
Experimental procedure
Figure 24 A schematic diagram of the system used for the TPD-NH3 analysis Note MFC =
mass flow controller
Approx 50 mg of a sample was transferred into a quartz tube reactor and placed into
an electrically heated furnace coupled with a temperature controller (Newtronic MICRO 96
TP10+) The sample was exposed to a helium flow (9999 BOC 50 mlmin) at 550 degC for 1
h and then cooled to 100degC (with liquid nitrogen) the temperature at which the composition of
the gas stream was adjusted to contain 10 wt of NH3 (9998 NH3 BOC) in He ie 100000
ppm NH3 The sample was exposed to the gas flow for 30 minutes and then rinsed with pure
105
He at the same temperature for 1 h in order to eliminate the species physically adsorbed on the
catalyst surface The desorption of ammonia was then started by increasing the temperature to
600 degC at a heating rate of 5 degCmin in a helium gas flow The amount of desorbed-NH3
detected by a mass spectrometer corresponds to the concentration of the acid sites while the
temperature at which ammonia desorbs correlates with the acid sitesrsquo strength The TPD plot
was logged using an in-line data acquisition system
The gas flow was measured and controlled by a mass flow controller (Brooks
Instrument 0152 Flow Control) Liquid nitrogen was supplied in a cryogenic vessel
(Statebourne Cryogenics Cryostor 60) The starting desorption temperature of 100 degC was used
based on the results reported by Hegde et al (1989) who carried out the TPD-NH3 analysis of
H-Beta zeolite starting at room temperature detecting large quantities of physisorbed NH3
molecules ie weakly NH3 bonded molecules not attached to the acid sites The TPD of NH3
starting at ge100 degC was adopted by several researchers (Katada et al 1997 Camiloti et al
1999 Rodriacuteguez-Gonzaacutelez et al 2008)
The concentration of acid sites was determined from the NH3 desorption profiles in
which the areas under the concentration against time curves were deconvoluted using the
OriginPro 86 (OriginLabreg) software of which the Gauss function was used for fitting multi-
peaks The absolute number of zeolite acid sites was determined by multiplying the area under
the peak by the total gas flow rate and the acidity was expressed in micromol g-1 of the catalyst
used
106
416 Thermogravimetric analysis (TGA) and Differential scanning
calorimetry (DSC)
Thermal analysis is a technique to measure specific physical properties of a substance
under controlled conditions of heating or cooling In practice thermal analysis is the heating
(or cooling) of a sample and benchmarking change with respect to an inert reference sample in
an identical thermal regime and recording any change of enthalpy
When a sample is heated any chemical change (eg decomposition) or physical change
(eg melting phase transition) is accompanied by exo- or endothermic peaks These are
recorded as differential thermal analysis (DTA) The DTA signal is the temperature difference
between sample and reference typically plotted versus temperature If mass changes also
occur these are recorded by thermogravimetry (TG) or differential thermogravimetry (DTG)
Similarly to DTA DSC is a technique measuring the amount of heat required to raise the
temperature of a sample and reference as a function of temperature In addition to temperature
it also measures the heat energy (eg heat of fusion) that is consumed by the sample
Thermogravimetric analysis data were measured using a TA Instruments TGA Q500
thermal analyser with the maximum operating temperature of 1000 degC Approx 15 mg sample
was placed in a flowing gas of N2 (400 mlmin) and air (600 mlmin) and heated up to 600 degC
at a rate of 10 degCmin This technique provided mainly information on the water content of the
sample and decomposition of salts DSC analysis was performed on a Differential Scanning
Calorimeter (DSC Q2000 V2410 Build 122) Approx 3 mg sample was placed in a flowing
gas of N2 (500 mlmin) and He (500 mlmin) and heated up to 500 degC at a rate of 10 degCmin
107
42 Characterisation of liquid and gaseous samples
Gas products were first collected through a valve cap of the tubular batch reactors using
a gas syringe and analysed by a mass spectrometer The liquids discharged from the reactor
were separated from solids by filtration using a syringe filter (Millex-MP 022 microm filter unit)
before being subjected to chemical analysis The identification and quantification of liquid
reaction products were performed by 1H-NMR utilising an internal standard (potassium
hydrogen phthalate KHP) of known-concentration added into each NMR sample using D2O
as a solvent The formula for conversion (mol) yield (C-mol) and selectivity (C-mol)
are given in Chapter 45 The 13C-NMR and correlation spectroscopy (COSY) were also used
to identify the reaction products A high-performance liquid chromatograph (HPLC) equipped
with an ion exchange column was employed to confirm the results using an external standard
method The quantitative analysis was performed after the construction of calibration curve(s)
Details of the equipment and analytical procedures used for the characterisation of gas and
liquid samples are described in Chapter 421 to 424
421 Nuclear magnetic resonance (NMR) spectroscopy
Theoretical background
The principle of NMR is based on the fact that nuclei of atoms have magnetic
properties and subatomic particles (ie protons electrons and neutrons) behave like magnetic
bars and spin (precession) around the nuclei (Skoog et al 2007) Quantum subatomic particles
(ie protons electrons and neutrons) have spin However in some atoms (such as 12C 16O 2H)
these spins are paired and cancel each other out so that the nucleus of the atom has no overall
spin
108
The nuclear magnetic resonance phenomenon occurs when the nuclei which initially
have their spins (for example +12 in the case of 1H) in parallel with the external (applied)
magnetic field are induced to absorb energy and subsequently excited into a higher energy
level at which their spins are anti-parallel (-12 in the case of 1H) (Skoog et al 2007) Once
application of the external magnetic field is stopped the nuclei emit energy and return to their
normal state This up and down movement of the nuclei is called ldquoresonancerdquo However in
order to allow the nuclei to resonate the energy (magnetic field having radio-frequency)
applied has to have the same frequency as the frequency spins of the nuclei The energy emitted
by the nuclei is detected by a radio frequency detector
Experimental procedure
In this work the NMR spectra were recorded on Bruker DRX-400 (1H 4001 MHz
13C 1006 MHz) and AV-400 (1H 4003 MHz 13C 1007 MHz) spectrometers at 298 degK
unless otherwise stated The NMR spectra were processed using MestreNova software 1H and
13C chemical shifts were reported as δ (in ppm) and referenced to the residual proton signal and
to the 13C signal of the deuterated solvent (1H D2O δ 470) respectively The following
abbreviations have been used for multiplicities s (singlet) d (doublet) t (triplet) m (unresolved
multiplet)
4211 Identification of liquid products
A typical 1H-NMR spectrum of liquid products of reactions performed under the
hydrothermal conditions studied (see Chapter 6 and 7) are shown in Figure 25 The peaks in
the 1H-NMR spectra were identified according to the chemical shifts listed in Table 17
109
Liquid products typically identified include acrolein acetol acetic acid lactic acid
ethanol methanol and acetaldehyde which exists in two different forms acetaldehyde and
acetaldehyde hydrate (11-dihydroxyethane) (Ahrens and Strehlow 1965) Traces of formic
acid were also detected but only in the reactions at either low temperature or short reaction
time
The quantification of liquid products was performed mainly by 1H-NMR due to
extensive overlapping of glycerol and lactic acid peaks by HPLC due to the
insufficient separation possibly due to the column
MeOH and EtOH were not detectable with the variable wavelength detector
(VWD) used and
inaccurate quantification of acetaldehyde hydrate with HPLC due to low R-
squared value of the calibration curve (constructed using acetaldehyde aqueous
solution)
110
Figure 25 1H-NMR spectrum of the reaction products detected after the glycerol dehydration
in the presence of 48 CeH-Beta at 330degC 1297 bars for 1 h Reaction conditions 01 M
glycerol (03 mmol) the weight ratio of glycerolcat = 1 the mole ratio of glycerol to acid sites
of the catalyst = 115
111
Table 17 The list of chemical shifts of reaction products detected by 1H-NMR after the
hydrothermal dehydration of 01 M glycerol (03 mmol) in the presence of 48 CeH-Beta at
330 degC 1297 bars for 1 h The weight ratio of glycerolcat = 1 the mole ratio of glycerol to
acid sites of the catalyst = 115
Products 1H NMR data (400 MHz D2O) ppm
acetic acid δ = 195 (1H s CH3COOH)
acetaldehyde δ = 954 (1H q CH3COH) 211 (3H d CH3COH)
acetaldehyde
hydrate
δ = 513 (1H q CH3CH(OH)2) 125 (3H d CH3CH(OH)2)
acetol δ = 203 (2H s CH3COCH2OH) 425 (3H s CH3COCH2OH)
acrolein δ = 936 (1H d CH2CHCOH) 656 (1H d CH2CHCOH) 642 (1H
d CH2CHCOH) 630 (1H m CH2CHCOH)
dihydroxyacetone δ= 430 (4H s (HOCH2)2CO)
ethanol δ = 104 (3H t CH3CH2OH) 350-356 (2H q CH3CH2OH)
formic acid δ = 822 (1H s HCOOH)
glycerol δ =344-358 (4H dd OHCH2CH(OH)CH2OH) 366-370 (1H m
OHCH2CH(OH)CH2OH)
glycolic acid δ = 403 (2H s HOCOCH2OH)
lactic acid δ = 130 (3H d CH3CH(OH)COOH) 423 (1H q
CH3CH(OH)COOH)
methanol δ= 323 (3H s CH3OH)
pyruvaldehyde δ= 125 (3H s CH(OH)2COCH3) 218 (CH(OH)2COCH3)
water δ= 471 (2H s H2O)
112
4212 Quantitative 1H-NMR analysis
The concentrations of feed glycerol and its liquid reaction products were determined by
quantitative 1H-NMR analysis and used for calculating the conversion (mol) of the feed
selectivity (C-mol) to and yield (C-mol) of liquid products The protocol consisted of the
following steps (i) to (iii)
(i) The NMR sample preparation
Each sample contained
150 μl of the reaction sample
150 μl of the internal standard solution (005 M or 01 M)
150 μl of deuterium dioxide (D2O)
Potassium hydrogen phthalate (KHP) was used as an internal standard for 1H-NMR
analysis throughout this work because it is solid non-hygroscopic and air-stable making it
easy to weigh accurately
A known amount of KHP was added into every 1H-NMR sample where its carboxylic
protons deprotonate in aqueous environment see Figure 26 so that only four aromatic protons
corresponding to the resonances at around 7-8 ppm appear in the spectra Although the signals
of the reaction products might also appear in this region only trace amounts were detected
113
Figure 26 Potassium hydrogen phthalate (KHP)
(ii) Quantitative analysis
The relative compound concentration was calculated from the area under its peak in 1H-
NMR spectra using the Equation 11 (Popor and Hallenga 1991)
Equation 11 The relationship between the concentration of a compound peak intensity and
the numbers of equivalent nuclei given the signals in 1H-NMR
CA
CB
= IA
IB
timesnB
nA
Where CA is the concentration of A
CB is the concentration of B
IA is the integrated intensity of A
IB is the integrated intensity of B
nA and nB are the numbers of equivalent nuclei given the signals
114
Example Calculation the concentration of lactic acid
The 1H-NMR spectra of lactic acid shows two peaks a doublet at 128 ppm and a
quartet at 423 ppm However only the doublet was used to evaluate the yield as it was not
overlapped by other peaks The two peaks of the internal standard were integrated calibrated
as 1 and then the area under the doublet found by integration was used for calculating the
relative concentration of lactic acid see Equation 12
Equation 12 The formula used for the calculation of relative concentration of lactic acid
Clactic acid= Ilactic acid
Istandard
times 2
3 timesCstandard
Clactic acid is the concentration of lactic acid produced
I lactic acid is the integrate intensity of the doublet peak of lactic acid
I standard is the integrate intensity of the internal standard
Cstandard is the concentration of the internal standard
nB is 2
nA is 3
each signal of the internal standard corresponds to 2 protons of the aromatic ring
the doublet of the lactic acid peaks correspond to 3 protons of the methyl group
(iii) Quantification calculation
The conversion yield (carbon basis) and selectivity (carbon basis) are defined in
Equation 13 14 and 15 respectively (Bicker et al 2005)
115
Equation 13 The formula used to find the conversion
where C0 is the initial and C1 the final concentration of the starting material (M)
Equation 14 The formula used to find the yield (mol - carbon basis)
Equation 15 The formula used to find the selectivity (mol - carbon basis)
4213 Validation of KHP
Ca(lactate)2 was used as a source of lactate to validate the concentration of the internal
standard KHP because it is solid air-stable non-hygroscopic making it easy to weigh
accurately Validation of KHP was performed by plotting the concentration of lactate obtained
from 1H-NMR against the theoretical values (see Figure 27) The linear regression equation
obtained indicates that the KHP is a suitable standard
116
Figure 27 Validation of KHP the internal standard for 1H-NMR
422 High-performance liquid chromatography (HPLC)
Theoretical background
HPLC is an analytical technique in which dissolved chemical species are separated due
to their different movement rates through a porous solid resin (known as a stationary phase)
inside a column A mixture of compounds is carried through the column by a mobile phase (a
solvent or a mixture of solvents) pumped through the column at high pressure and a fixed flow
rate Species with weak attraction to the stationary phase are eluted from the column first and
subsequently detected by a detector The UV absorption intensity of chemical species is
recorded against the retention time allowing both qualitative and quantitative analysis HPLC
y = 1155x + 04549
Rsup2 = 09953
0
2
4
6
8
10
12
14
16
18
0 2 4 6 8 10 12 14 16
[lact
ate
] exp
erim
en
t (m
M)
[lactate]theoritical (mM)
117
technique is suitable for low volatile- or high molecular weight organic compounds High
pressure makes the separation much faster than simple column chromatography
Experimental procedure
HPLC analyses were carried out using a complete Hewlett Packard 1050 HPLC system
equipped with a variable wavelength detector (VWD) and an ion exchange column (Aminex
HPX-87H 30cm x 78mm ID Bio-rad) The system was controlled by the computer software
HP Chemstation The VWD wavelength (λ) used 190 nm the column temperature 60 degC the
mobile phase 3 mM H2SO4 flow rate 05 mlmin
HPLC analysis was employed to confirm the analysis results obtained by NMR A
typical HPLC chromatogram of liquid samples of the reactions performed under the
hydrothermal conditions studied (see Chapter 6 and 7) is shown in Figure 113 in the Appendix
The calibration curves for glycerol and its reaction products were constructed using
commercial analytical standards The concentrations of unknown samples were determined
using a linear regression equation The retention times of the standards are listed in Table 18
while the calibration curves are provided in Figure 114 115 and 116 (in the Appendix)
Table 18 The retention times of compounds detected by Aminex 87H conditions 3mM
H2SO4 VWD 190 nm column temperature 60 degC flow rate 05mlmin
Retention time
(min)
Compounds
Retention time
(min)
Compounds
13548 pyruvaldehyde 17665 acetic acid
14233 glycolic acid 20322 acetol
14597 lactic acid 20846 acetaldehyde1
15264 glycerol 22276 acetaldehyde hydrate
15590 DHA 26502 acetaldehyde 2
16114 formic acid 30760 acrolein
118
4221 Quantitative analysis of results from HPLC vs 1H-NMR
The results from HPLC were close to those obtained from 1H-NMR (see Table 19) with
a difference of 6 However as the HPLC analysis was performed at a later date as compared
to 1H-NMR there is a possibility that some compounds degraded contributing to the difference
Several peaks in the HPLC chromatogram (especially at 42335 and 66766 min) remained
unidentified The GC-MS analysis was performed on selected samples where necessary
Table 19 Summary of residual glycerol concentrations quantified by 1H-NMR and HPLC The
sample was obtained from hydrothermolysis of glycerol in the presence of H-Beta zeolite at
270 degC 55 bar 3 h the glycerolzeolite weight ratio = 1 the mole ratio of glycerol to acid sites
of H-Beta zeolite = 451
Concentration of Glycerol determined by 1H-NMR Concentration of Glycerol determined by HPLC
00302 mM 00284 mM
423 Inductively Coupled Plasma - Optical Emission Spectrometry
(ICP-OES)
Theoretical background
ICP-OES is an analytical technique used for the detection of a broad range of elements
It is a type of emission spectroscopy that uses the inductively coupled plasma at high
temperature up to 9727 degC to produce excited or ionized atoms that emit electromagnetic
radiation at wavelengths characteristic of a particular element upon returning to ground state
119
(Skoog et al 2007) The emitted waves are measured by an optical spectrometer A detector
converts light energy (photons) from analyte emissions generated in the plasma into an
electrical signal that can be quantified A schematic of an ICP-OES is depicted in Figure 28
Figure 28 Schematic of an ICP-OES The figure was taken from (Verheijen 2012)
Experimental procedure
The liquids recovered from the reactors were filtered using a 02 microm syringe filter
diluted twenty times with 20 M HNO3 and analysed for aqueous Al Si and Cu using ICP-OES
(Optima 2000DV Perkin Elmer) The standard solutions (1000 ppm) were purchased from
MBH and Sigma Aldrich
424 Mass spectrometry (MS)
Mass spectrometry is a versatile analytical technique widely used for both qualitative
and quantitative chemical analysis of species present in a sample by measuring the mass-to-
charge ratio and abundance of gas-phase ions
120
Theoretical background
A mass spectrometric analysis involves the following steps (Skoog et al 2007) (1)
ionisation of the atom or molecule to form a positively charged ion (2) acceleration of the ions
to the same kinetic energy (3) deflection of the ions in a magnetic field based on their mass-
to-charge ratio (mz) where m is the mass number of the ion in atomic mass units and z is the
number of fundamental charges that it bears and (4) the detection of ions A mass spectrum is
a plot of the ion signal intensity as a function of the mass-to-charge ratio
Gaseous samples were analysed using a mass spectrometer (European Spectrometry
System II)
425 UV-VIS spectrophotometry
UV-VIS spectrophotometry was used to quantify the final H2O2 concentration of the
mother liqueurs recovered after the oxidative hydrothermolysis of glycerol using H2O2 This
method makes use of a chemical reaction between TiOSO4 and H2O2 producing yellow colour
pertitanic acid (Equation 16) stable for at least 6 h (Eisenberg 1943) see Figure 29 allowing
the spectrophotometry to be carried out at later time Moreover TiOSO4 selectively reacts with
H2O2 and is neither affected by other oxidizing agents nor by air (Amin and Olson 1967)
Equation 16 Reaction between TiOSO4 and H2O2 in water producing H2TiO4 (pertitanic acid)
1198791198944+ + 11986721198742 + 21198672119874 rarr 11986721198791198941198744 + 4119867+
121
Figure 29 Calibration solutions with increasing H2O2 concentrations after the reaction with
TiOSO4
Beckman Coulter DU-640 UV-VIS spectrophotometer was used for the measurements
The wavelength for pertitanic acid determination was 409 nm H2O2 calibration solutions were
prepared from 03 M (300000 μmolL) H2O2 stock solution A nearly linear calibration curve
was obtained in the range of H2O2 concentrations from 100 μmolL to 2000 μmolL The
concentrations of H2O2 in the samples were determined using a linear regression equation
122
Figure 30 A calibration curve of H2O2 solutions obtained using the titanium method
(Eisenberg 1943) the formed pertitanic acid was determined by UVVis spectrophotometry at
409 nm
4251 Preparation of TiOSO4 solution
TiOSO4 solution was prepared using a method adapted from Eisenberg (1943) 25 g
TiOSO4 (29 Ti as TiO2) was dissolved in 1 L of a 20 M H2SO4 solution continuously cooled
in an ice-water bath The solution was then stored in a fridge for up to three months The final
concentration of Ti (as TiO2) in the solution was ge 7480 ppm
4252 Preparation of H2O2 stock solution
Each H2O2 calibration solution was prepared by the dilution of 03 M H2O2 stock
solution the H2O2 concentration of which was determined by redox titration with KMnO4
y = 00005x - 00158
Rsup2 = 09995
0
02
04
06
08
1
12
0 500 1000 1500 2000 2500
Ab
sro
ba
nce
H2O2 (micromolL)
123
03 M H2O2 aqueous solution was prepared as follows 34015 g (003 mol) of 30 wt H2O2
solution was transferred into a 100 ml volumetric flask and the volume was adjusted to 100 ml
with DI water The solution was standardized with 00644 M KMnO4 solution (10230 g of
995 KMnO4 in 100 ml DI water) In a typical redox titration 3 g of conc H2SO4 (95) was
added slowly into 20 ml of 03 M H2O2 The acidified solution was then titrated with the 00644
M KMnO4 standard solution The balanced chemical equation for the overall reaction between
H2O2 and permanganate (1198721198991198744minus) is shown in Equation 17
Equation 17 The reaction between H2O2 and KMnO4 in an acidic solution (Jeffery et al 1989)
511986721198742 + 21198701198721198991198744 + 311986721198781198744 rarr 21198721198991198781198744 + 11987021198781198744 + 81198672119874 + 51198742
H2O2 which is a strong oxidant becomes a reducing agent in the presence of an even
stronger oxidizing agent such as permanganate H2O2 reduces permanganate (Mn7+) which is
intense purple in colour to a colourless product of manganese salt (Mn2+) The reaction occurs
in an H2SO4 solution as protons are required to form water
124
43 Chemicals and materials
Unless otherwise stated all chemicals used were of standard reagent grade and were
used without further purification Table 20 lists all the chemicals and materials used in present
work
Table 20 The list of chemicals and materials used in the present work
Chemical reagents Supplier
Ce(NO3)36H2O 99999 trace metals basis Sigma Aldrich
CeO2 999 trace metals basis Sigma Aldrich
Cu(NO3)23H2O analytical reagents 99-104 Sigma Aldrich
CuO powder 990 Sigma Aldrich
LaCl37H2O ACS reagent 645-700 Sigma Aldrich
La(NO3)3middotH2O analytical reagents ge990 (titration) Sigma Aldrich
SnCl22H2O ACS reagent ge98 Sigma Aldrich
Zn(SO4)27H2O ReagentPlusreg ge990 Sigma Aldrich
triethanolamine (TEA) (97 wt) ACROS Organicstrade
triethylammoniumhydroxide (TEAOH) (40wt) Sigma Aldrich
tetraethylorthosilicate (TEOS) (999) reagent Sigma Aldrich
sodium aluminate anhydrous (NaAlO2) technical Sigma Aldrich
bentonite clay (montmorillonite Al2O3middot4SiO2middotH2O) Sigma Aldrich
methyl cellulose (viscosity 4000cP 2 in H2O) Sigma Aldrich
γ-alumina 9997 trace metals basis Alfa Aesar
NH4-ZSM-5 (CBV2314) Zeolyst International USA
NH4-Beta zeolite (CP814E) Zeolyst International USA
13-propanediol 98 wt Sigma Aldrich
glycerol gt 99 wt Sigma Aldrich
12-propanediol ACS reagent ge995 Sigma Aldrich
glyceraldehyde ge90 (GC) Sigma Aldrich
pyruvaldehyde 40 wt in H2O Sigma Aldrich
lactic acid 88-92 Sigma Aldrich
CeCl37H2O 99999 trace metals basis Sigma Aldrich
ZnCl2 99999 trace metals basis Sigma Aldrich
125
Table 20 (continued)
Chemical reagents Supplier
formic acid 98 Sigma Aldrich
acetic acid glacial ge9985 Sigma Aldrich
acetaldehyde 995 Sigma Aldrich
acrolein 95 Sigma Aldrich
methanol 99 VWR
ethanol absolute 9998 VWR
potassium hydrogenphthalate (KHP) BioXtra ge9995 Sigma Aldrich
deuterium oxide (D2O) 999 atom D Sigma Aldrich
ICP multi-element standard solution IV Merck
MBH analytical Multi-element plasma standard 3 Element
Custom Multi
MBH
Cu (1000mgL) ICP standard solution TraceCERTreg Sigma Aldrich
Al (1000mgL) ICP standard solution TraceCERTreg Sigma Aldrich
H2O2 (30 wv) VWR
H2SO4 95 VWR
HNO3 68 VWR
KMnO4 995 BDH Analar
titanium oxysulfate (TiOSO4 ge29 TiO2) technical Sigma Aldrich
silica beads (particle size ~1 mm) Assistant Germany
carborundum powder (silicon carbide) (60 Grit) T N Lawrence Art Supplies
acetol 90 wt Sigma Aldrich
pyruvic acid 98 wt Sigma Aldrich
Sn(CH3COO)2 Sigma Aldrich
Zn(CH3COO)2 Sigma Aldrich
glycolic acid ReagentPlusreg 99 Sigma Aldrich
13-dihydroxyacetone dimer 97 Sigma Aldrich
Amberlitereg IR120 hydrogen form Sigma Aldrich
126
44 Catalyst preparation
A series of catalysts based on commercial NH4-Beta zeolite or NH4-ZSM-5 was
prepared following the steps outlined in Figure 31 All metal-doped zeolites prepared using
solid state ion exchange reaction (SSIE) with metal salts at 550 degC are listed in Table 21 and
22 The list of sample names of powdered zeolite prepared and tested as catalysts in tubular
batch reactors is given in Table 23 The names are composed of the weight percentage of metal
relative to the weight of zeolite For instance 48 CeH-Beta zeolite stands for 48 wt Ce
content relative to H-Beta zeolite
127
Figure 31 The catalysts preparation based on NH4-Beta zeolite deAl-Beta refers to
dealuminated H-Beta zeolite Identical steps except for the extrusion and the preparation of
metal-doped deAl-ZSM-5 were used for the catalyst preparation based on NH4-ZSM-5 Metal
salts used as the source of corresponding metal ions included Ce(NO3)3middot6H2O La(NO3)3middotH2O
Sn(CH3COO)2 Zn(CH3COO)2 Cu(NO3)23H2O
128
Table 21 A summary of all metal-doped zeolites based on commercial Beta zeolite prepared
in this work The solid state ion exchange (SSIE) technique was used utilizing metal nitrates
as the source of metal ions The metal-doped zeolites were used for trials with tubular batch
reactors
Zeolite
wt of metal content relative to the zeolite
- Ce Ce La Sn Zn Cu
0 25 48 48 48 48 48
NH4-Beta
H-Beta
deAl-Beta
deAl-Beta refers to dealuminated H-Beta zeolite
Table 22 A summary of all metal-doped zeolites based on commercial ZSM-5 zeolite
prepared in this work The solid state ion exchange (SSIE) technique was used utilizing metal
nitrates as the source of metal ions The metal-doped zeolites were used for trials with tubular
batch reactors
Zeolite
wt of metal content relative to the zeolite
- Ce Ce La Sn Zn Cu
0 25 48 48 48 48 48
NH4-ZSM-5
H-ZSM-5
deAl-ZSM-5
deAl-ZSM-5 refers to dealuminated H-ZSM-5
129
Table 23 The list of sample names of powdered zeolite catalysts prepared and tested as catalyst
in tubular batch reactors
Catalyst Catalyst
NH4-Beta NH4-ZSM-5
H-Beta zeolite H-ZSM-5
48 CeNH4-Beta 48 CeNH4-ZSM-5
48 LaNH4-Beta 48 LaNH4-ZSM-5
48 SnNH4-Beta 48 SnNH4-ZSM-5
48 ZnNH4-Beta 48 ZnNH4-ZSM-5
48 CeH-Beta 48 CeH-ZSM-5
48 LaH-Beta 48 LaH-ZSM-5
48 SnH-Beta 48 SnH-ZSM-5
48 ZnH-Beta 48 ZnH-ZSM-5
48 CuH-Beta Dealuminated H-ZSM-5 (deAl-ZSM-5)
Dealuminated H-Beta zeolite (deAl-Beta)
1 CedeAl-HBetaII4h
2 CedeAl-HBetaII4h
3 CedeAl-HBetaII4h
48 CedeAl-BetaII4h
48 CudeAl-BetaII4h
48 SndeAl-BetaII4h
48 ZndeAl-BetaII4h
the Roman and Arabic numerals indicate the number of grams of H-Beta zeolite dealuminated
in 20 ml of 13 M HNO3 and the reaction time in hour (ie 4 h) respectively
The finely powdered zeolites prepared (Table 21 and 22) were tested as catalysts in tubular
batch reactors However they were unsuitable for use in a continuous flow reactor due to their
tendency to get washed out The issue was addressed by attempts to immobilise the zeolites by
the following means
(i) synthesis of CeH-Beta zeolites on hollow fibre ceramic membranes
130
(ii) incorporation of Beta zeolite-based catalysts into mesoporous silica matrix
followed by pelletization
(iii) extrusion of Beta zeolite-based catalysts with clay (bentonite) or γ-alumina
See Table 24 for a summary of all zeolite-based catalysts prepared for the trials with a
continuous flow reactor in the present work Experimental protocols for the preparation of all
the catalysts are presented in Chapter 441 to 446
Table 24 A summary of all zeolite-based catalysts used for the trials with a continuous flow
reactor in the present work
Production method Catalyst
H-Beta 48 CuH-Beta 48 CeH-Beta
Synthesis on ceramic membrane support
Incorporation into mesoporous silica matrix
(TUD-1)
Extrusion with clay
Extrusion with alumina
441 Preparation of H-Beta zeolite and H-ZSM-5
20 g of NH4-Beta zeolite (SiO2Al2O3 = 25) or NH4-ZSM-5 (SiO2Al2O3 = 23) were
activated by calcination at 550 degC for 5 h in air atmosphere in a muffle furnace (Elite Thermal
Systems Limited) at a heatingcooling rate of 100 degCh to form the acidic form of zeolites (H-
zeolites) see Equation 18 The products were characterised by XRD BET-N2 and TPD-NH3
See Chapter 51 for the results
131
Equation 18 Calcination of NH4-zeolites to form H-zeolites
442 Preparation of metal-exchanged NH4-Beta H-Beta NH4-
ZSM-5 and H-ZSM-5 zeolites
The metal-doped zeolite catalysts were prepared via a solid state ion exchange (SSIE)
reaction between (i) the as received zeolites (NH4-ZSM-5 or NH4-Beta zeolite) or activated
zeolites (H-ZSM-5 or H-Beta zeolite) prepared according to the experimental protocol
described in Chapter 441 and (ii) a desired metal salt Ce(NO3)3middot6H2O La(NO3)3middotH2O
Sn(CH3COO)2 Zn(CH3COO)2 Cu(NO3)23H2O The zeolite doped with 48 wt of metal was
prepared as follows 2 g of NH4- or H- Beta zeolite (or NH4- or H- ZSM-5) were mixed with a
required amount of metal salt and manually ground in a ceramic mortar for 15 min at ambient
conditions The mixture was then transferred into an alumina crucible and calcined at 550 degC
for 3 h in air atmosphere using a heatingcooling rate of 100 degCh The calcined samples were
labelled as ldquo48 MH-zeoliterdquo or ldquo48 MNH4-zeoliterdquo where the ldquo48rdquo indicates the
weight percentage of metal relative to the weight of zeolite and the letter ldquoMrdquo denotes the type
of metal ion used The samples were characterised by XRD TPD-NH3 TEM and SEM See
Chapter 52 for the results
132
443 Preparation of metal-exchanged dealuminated H-Beta zeolite
The preparation of metal-doped dealuminated H-zeolite (H-Beta or H-ZSM-5) method
consists of three steps
1 Preparation of H-zeolite
2 Dealumination of H-zeolite
3 Doping of dealuminated H-zeolite with host metal ions
The H-zeolite was prepared following the description in Chapter 441 The
dealumination of the sample was then performed using the method reported by Hammond et
al (2012) This method is capable of removing Al atoms without collapsing the crystalline
structure of the zeolite
The dealumination was carried out as follows 125 g of H-zeolite was dispersed in 25
ml of 13 M HNO3 (20 mlg zeolite) sealed in a 50 ml Teflon lined stainless steel autoclave
reactor and heated at 100 degC for 1 4 or 20 h in a thermostatically controlled air recirculating
oven (GC oven HP Hewlett Packard 5890) The reactor was subsequently cooled down to
room temperature (RT) the reaction liquid was decanted and the solid re-suspended in 100 ml
of deionised water The sample was then centrifuged at 3000 rpm (Heraeus-Megafuge 40R
Centrifuge Thermo Scientific) for 20 min and decanted with the cycle repeating three times
after when the pH of the sample reached ~7 The recovered solid was then oven-dried for 12 h
at 110 degC and stored in a sealed container at RT (Lami et al 1993) The prepared solids were
characterised by XRD TEM XRF and NH3-TPD
The desired metal ions were then incorporated into a dealuminated zeolite by SSIE
method according to the protocol in Chapter 442 The metal salts used as precursors included
Ce(NO3)3middot6H2O La(NO3)3middotH2O Sn(CH3COO)2 Zn(CH3COO)2 CeCl37H2O SnCl22H2O
133
and ZnCl2 The two reaction steps of the dealumination of H-zeolites followed by SSIE is
shown in Equation 19
Equation 19 Preparation of metal-doped dealuminated H-zeolites A treatment with conc
HNO3 creates vacant sites (in the zeolite framework) which were subsequently filled with host
metal atoms supplied by a metal salt
444 Preparation of CeH-Beta zeolite on ceramic hollow fibre
membranes
The CeH-Beta zeolite on a ceramic hollow fibre membrane (CHFM) support was
prepared using a method adapted from Kantam et al (2005) Yadav et al (2009) and Yang et
al (2010) The procedure consisted of the following steps
1 sol-gel synthesis of Na-Beta zeolite on CHFM
2 conversion of Na-Beta zeolite on CHFM to NH4-Beta zeolite
3 conversion of NH4-Beta zeolite on CHFM to H-Beta zeolite
4 conversion of H-Beta zeolite on CHFM to CeH-Beta zeolite
134
1 Sol-gel synthesis of Na-Beta zeolite on CHFM
Solution A Water (27891 g) was added slowly into TEOS (45355 g 2177 mmol) and stirred
for 30 min
Solution B Water (2 g) was added into TEAOH (48195 g 1310 mmol) followed by slow
addition of NaAlO2 (00419 g 050 mmol) The solution mixture was stirred for 30 min
The CHFM were provided by Dr Zhentao Wu of the department of Chemical
Engineering Imperial College London (Kingsbury et al 2010) The alumina hollow fibre
precursors received were cut into approx 5 mm long pellets and calcined in a muffle furnace
(Eurotherm 2116 Elite Thermal Systems Ltd) to yield CHFM The temperature was increased
from room temperature (RT) to 600 degC at 2 degCmin and held for 2 h and then to 1450 degC at 5
degCmin and held for 4 h The temperature was then reduced to RT at a rate of 3 degCmin
Figure 32 Alumina hollow fibre precursors of ceramic hollow fibre membranes used as a
support for CeH-Beta zeolite in the present work
135
Na-Beta zeolite on CHFM was synthesized as follows 2 g of CHFM pre-dried in a
vacuum oven for 1 h at RT was added into solution A and stirred for 30 min at 500 rpm The
solution B was then added dropwise into the stirred mixture producing a white precipitate The
mixture was stirred for another 2 h The product was then sealed in a 50 ml Teflon-lined
autoclave reactor placed in a thermostatically controlled air recirculating oven (HP Hewlett
Packard 5890) at 160 degC for 48 h and subsequently cooled down by an ice-water bath The
mother liquor was decanted and the solid dispersed in 100 ml DI water and centrifuged at 3000
rpm for 20 min at RT The cycle was repeated three times The recovered solid containing a
mix of zeolite and CHFM pellets was oven-dried at 110 degC for 12 h
2 Conversion of Na-Beta zeolite on CHFM to NH4-Beta zeolite
NH4-Beta zeolite on CHFM was prepared by liquid ion exchange reaction of the Na-
Beta zeolite on CHFM obtained in step 1 with 2 M aqueous solution of NH4Cl 2 g of the mix
of Na-Beta zeolite and CHFM was stirred with 120 ml of 2 M NH4Cl(aq) at 80 degC for 6 h The
suspension was filtered on stainless steel mesh with 05 mm aperture diameter and washed
three times with 50 ml of DI water prior to drying at 110 degC for 12 h The recovered filtrate
was diluted with 100 ml DI water and centrifuged at 3000 rpm for 20 min to recover the
suspended zeolite The solid was dried at 110 degC for 12 h and stored at RT To ensure a
complete ion exchange the whole cycle of the liquid ion exchange was repeated three times
3 Conversion of NH4-Beta zeolite on CHFM to H-Beta zeolite
NH4-Beta zeolite on CHFM prepared in step 2 was transformed into H-Beta zeolite by
calcination at 550 degC for 5 h at a heatingcooling rate of 100 degCh The NH4-Beta zeolite
136
recovered from the filtrate in step 2 was converted into its H-Beta form using identical
procedure and used as a control sample
4 Conversion of H-Beta zeolite on CHFM to CeH-Beta zeolite
The samples of (i) H-Beta zeolite on CHFM and (ii) control H-Beta zeolite powder
were transformed into their CeH-Beta form by a liquid ion exchange reaction with soluble
cerium salt 2 g of sample was stirred with 120 ml of 2 M Ce(NO3)3 at 80 degC for 6 h The work-
up procedure is identical to that described in step 2 The obtained CeH-Beta zeolite powder
and CeH-Beta zeolite on CHFM were characterised by XRD and SEM see Chapter 551
445 Preparation of H-Beta and metal-dopedH-Beta zeolites on
TUD-1 matrix
To agglomerate the fine particles of zeolite catalyst the H-Beta zeolite (or 48 CeH-
Beta or 48CuH-Beta) was incorporated into the pores of mesoporous silica matrix TUD-1
using a method based on the work of Lima et al (2010) Using this procedure the final product
consisted of 40 wt of the metal-doped Beta zeolite and 60 wt TUD-1
1587 g (1032 mmol) of TEA was mixed with 12 g of DI water at RT for 15 min 4 g
of H-Beta zeolite (or 48 CeH-Beta or 48 CuH-Beta) was then slowly admixed and the
mixture was stirred for 30 min to ensure complete dispersion 2136 g (1024 mmol) of TEOS
was then added dropwise into the mixture to allow the aggregation of silica particles to occur
gradually The mixture was stirred for 2 h after which 120752 g (328mmol) of TEAOH was
added dropwise and the solution was stirred for another 2 h The resulting light-brown colour
137
gel was left to age under static conditions for 24 h at RT It was dried at 100 degC for 24 h and
then transferred into a 50 ml Teflon line autoclave reactor heated at 180 degC for 8 h followed
by rapid cooling in an ice-water bath The brown colour gel was calcined at 600 degC for 10 h at
a heating rate of 1 degCmin The resulting white powder containing a mixture of TUD-1 and
zeolite was analysed by XRD and BET-N2 see Chapter 552 The powder sample was pressed
into a pellet using a force of 50kN for 10 min crushed using a pestle and mortar and sieved
The particles with sizes ranging from 250-450 μm were used in a packed bed reactor
446 Preparation of extruded zeolite catalysts
The extrusion of H-Beta zeolite and 48 CuH-Beta was performed using two types
of binders bentonite clay and γ-Al2O3
Extrusion with bentonite binder The experimental protocol for the preparation of extruded
zeolite catalysts using bentonite as a binder was adapted from Melero et al (2012) 65 g H-
Beta zeolite 3 g bentonite and 05 g methyl cellulose (MC) were manually mixed together 20
g of DI water was then added and mixed until all the material became a paste The material
was then transferred into a 10 ml plastic syringe used as an extruder The extruded catalyst rods
were left to air dry for 15 min and cut into approx 2 mm long pellets see Figure 33A To
slowly dry the material the pellets were transferred on a metal sieve placed above the water
level inside a 25 L crystallization dish and covered with a lid (see Figure 33B) prior to leaving
it at 20 degC for 24 h The pellets were left to dry at 20degC on the work bench for another 24 h
followed by the calcination at 550degC using a heating rate of 03degCmin
138
Figure 33 The extruded catalyst rods were cut into approx 2 mm long pellets (A) which were
then dried slowly in humid environment at RT (B)
Extrusion with γ-Al2O3 binder The experimental protocol for the extrusion of 48 CuH-
Beta zeolite using γ-Al2O3 (lt106 microm) as a binder was adapted from Sntamaaria et al (2002)
Ramiacuterez et al (2005) and Kasture et al (2007) To prepare a paste of 6040 weight ratio of
zeolite to binder 24 g of 48 CuH-Beta zeolite and 16 g of γ-Al2O3 were manually mixed
together in a 100 ml plastic beaker 5 ml of 3 wt acetic acid aqueous solution was then added
and mixed until the material became a homogeneous paste The paste was then transferred into
a 10 ml plastic syringe used as an extruder The extruded catalyst rods were left to air dry for
5 min and cut into approx 2 mm long pellets The prepared pellets were stored in open
atmosphere at RT overnight followed by drying at 40 degC for 2 h and then 2 h at 120 degC Finally
the pellets were calcined at 550 degC for 4 h at a heating rate of 5 degCmin
The original method used by Sntamaaria et al (2002) Ramiacuterez et al (2005) and
Kasture et al (2007) was modified by the addition of MC in order to improve the consistency
139
of the extruded paste (Mitchell et al 2013) The paste consisted of 575 wt H-Beta zeolite
(or 575 wt of 48 CuH-Beta zeolite) 375 wt γ-Al2O3 and 5 wt MC where the
numbers indicate the weight percentage of each component relative to the total weight of the
mixture 65 ml of 3 wt acetic acid aqueous solution was required to prepare the paste of H-
Beta zeolites while the paste of 48 CuH-Beta zeolite needed 55 ml The extrusion protocol
was identical to that for the preparation of extruded zeolite catalysts using alumina as a binder
140
45 Hydrothermal dehydration of glycerol
451 Dehydration of glycerol
Hydrothermal transformations of glycerol was investigated using tubular batch reactors
heated in a thermostatically controlled air recirculating oven (HP Hewlett Packard 5890) see
Figure 34 The cylindrical tube reactors were made of stainless steel (SS316 Ham-Let Let-
Lokreg dimension internal diameter (id) ~085 cm length ~1063 cm volume ~60 cm3) and
capped at both ends where one end was a valve cap A control reactor was fitted with a K-type
thermocouple placed inside the reactor at ~5 cm length and connected to a digital thermometer
(Comark KM340)
The empty reactors were weighted filled with 3 ml of 01 M stock solution of glycerol
(09209 g (001 mol) of glycerol (gt99 wt) in 100 g DI water) and a pre-determined amount
of a catalyst if required The reactors were purged with N2 gas for 5 min and sealed and the
final weight of the reactors were recorded Each reactor was then placed in an oven at RT and
heated to 270 degC 300degC or 330 degC for the desired period of time The time required for heating
up the reactors to an operating temperature ranged between 8-15 min and was not included in
the recorded reaction times The reactors were subsequently transferred into an ice-water bath
to stop the reactions dried with acetone and then reweighed to check for possible leakage
Multiple reactors were often heated simultaneously and withdrawn individually at regular time
intervals Finally gaseous products were collected through a valve cap of the reactor and the
reactor contents were left to settle The solution was then filtered using a 022 microm nylon syringe
filter while the recovered solids were dried in an oven at 65 degC for 16 h The solid samples
were analysed by XRD spectroscopy while the liquid samples were analysed by 1H-NMR
HPLC and ICP-OES Gaseous products were analysed by mass spectrometry The operating
141
pressures were calculated as a sum of partial pressures of saturated vapour pressures of glycerol
and water at the operating temperatures using the Aspen plus V84 software The calculated
pressures were verified using a separate batch reactor equiped with a thermocouple and
pressure gague
Figure 34 The thermostatically controlled air recirculating oven (GC HP Hewlett Packard
5890) containing a set of bomb reactors
452 Dissolution of H-Beta zeolite in hot liquid water
The dissolution of H-Beta zeolite in hot liquid water (HLW) was investigated using the
equipment listed in Chapter 451 The reactors of known initial weights were filled with 30
mg of H-Beta zeolite and 3 ml of DI water purged with N2 gas for 5 min sealed and the final
weights were recorded Each reactor was placed in an oven at RT and heated to 150 200 and
250 degC for 1 3 5 and 20 h The reactors were subsequently transferred into an ice-water bath
to stop the reactions dried with acetone and reweighed to check for possible leakage Finally
142
the reactor contents were left to settle at RT decanted and filtered using a 022 microm nylon
syringe filter The recovered solids were dried in an oven at 65 degC for 16 h and analysed by
XRD The liquids were analysed by ICP-OES The dissolution of H-Beta zeolite in water at 20
degC (RT) for 1 24 and 48 h was performed in a closed vial without stirring 30 mg of H-Beta
zeolite and 3 ml of DI water were added in a 5 ml glass vial they were then hand shaken for
one minute and left at RT The solutions were filtered after 1 24 and 48 h and analysed by the
same manner as the samples prepared at 150 200 and 250 degC but analysis of solids was not
carried out
453 Effect of dissolved Al and Si on the dehydration of glycerol
To evaluate the effect of aqueous Al and Si species on the hydrothermal dehydration of
glycerol in this work a two-step experiment was carried out
Step 1 ndash Preparation of dissolved aqueous Si and Al species
The experiment was performed using the equipment listed in Chapter 451 The
reactors of known initial weights were filled with
3 ml of DI water and
30 mg H-Beta zeolite or 30 mg dealuminated Beta zeolite (deAl-BetaI20h)
The reactors were purged with N2 gas for 5 min sealed and the final weights were
recorded Each reactor was placed in an oven at RT and heated at 330 degC under autogenic
pressure for 5 10 15 30 45 and 60 min The reactors were subsequently transferred into an
ice-water bath to stop the reactions dried with acetone and reweighed to check for possible
143
leakage Finally the resulting solutions were filtered at RT using a 022 microm nylon syringe filter
and then analysed by ICP-OES
Step 2 - Hydrothermal dehydration of glycerol in zeolite filtrates
The experiment was performed using the equipment described in Chapter 451 3 ml
of the filtered solutions obtained in Step 1 further referred to as zeolite filtrates were
transferred into a tube reactor together with 30 mg of 99 glycerol purged with N2 for 5 min
sealed and heated at 330 degC for 5 10 15 30 45 or 60 min The holding time used was identical
to that used for preparing the zeolite filtrate for instance the 5 min dehydration of glycerol
was performed in the zeolite filtrate prepared with the holding time of 5 min Once the required
times were reached the reactors were cooled down in an ice-water bath the liquids were
filtered analysed by ICP-OES and 1H-NMR
46 Hydrothermal oxidation of glycerol
461 Batch process
The initial work focused on finding of the suitable reaction conditions for the
hydrothermal oxidation of glycerol using tubular batch reactors for the simplicity of use The
equipment and the experimental procedure used are described in Chapter 451 However the
feed stock solution catalysts and reaction conditions used for the oxidation were different and
are provided in this chapter
The empty reactors were weighted filled with 3 ml of a mixture of 02 M aqueous
solution of glycerol and 02 M H2O2 and a pre-determined amount of a catalyst if required
The reactors were purged with N2 gas for 5 min sealed and the final weight of the reactors
144
were recorded Each reactor was then placed in an oven at RT and heated to 125 150 and 175
degC under autogenic pressure (water and H2O2) for the desired period of time The reactors were
subsequently transferred into an ice-water bath to stop the reactions dried with acetone and
reweighed to check for possible leakage Finally the reaction solution was left to settle at RT
the solution was decanted and filtered using a 022 microm nylon syringe filter The recovered
solids were dried in an oven at 65 degC for 16 h The solids were analysed by XRD while the
liquids by 1H-NMR HPLC ICP-OES and UV-VIS
The feed stock solution used was a mixture of 02 M aqueous solution of glycerol and
02 M H2O2 prepared as follows 18418 g (002 mol) of glycerol (gt99 wt) and 22676 g (002
mol) of H2O2 (30 wv) were transferred into a 100 ml volumetric flask and the volume was
adjusted to 100 ml with DI water
462 Continuous flow process
A continuous flow system assembled according to the diagram in Figure 35 was used
to investigate the hydrothermal oxidation of glycerol in subcritical water A photo of the
apparatus is shown in Figure 36
145
Figure 35 A schematic of the high pressure continuous flow system apparatus used for glycerol oxidation under sub-CW conditions
146
Figure 36 The high pressure continuous flow system apparatus used for investigation of
glycerol oxidation under sub-CW conditions A HPLC pump no 1 for preheated water a DI
water reservoir B HPLC pump no 2 for feed solution b feed reservoir C preheater D
tubular reactor E digital thermometer ersquo K-type thermocouple inside the reactor erdquo K-type
thermocouple inside the copper block F electrical furnace G furnace temperature controller
H digital thermometer at the joining point between the feed solution and pre-heated water I
cooling coiled tube in an ice-water bath J stainless steel in-line filter K pressure relief valve
L digital pressure gauge M back pressure regulator N gas-liquid separator O liquid sample
P low pressure gas tube leading to a fume extractor
I High pressure continuous flow system apparatus
Two cylindrical stainless steel tubes (SS316 Ham-Let Let-Lokreg) were fabricated to
the following specifications and used as
147
(i) a flow through water preheater - id ~165 mm length ~500 mm volume ~11 cm3 and
(ii) a plug flow or packed bed reactor - id ~437 mm length ~500 mm volume ~75 cm3
Both the reactor [D] and preheater [C] are housed within a copper block (length 40 cm)
placed in a vertical tube furnace [F] (vertical length 40 cm) with the maximum operating
temperature of 1100 degC see Figure 35 and 36 The temperature of the furnace is controlled by
a digital furnace controller [G] The copper block is fitted with a K-type thermocouple [ersquorsquo]
placed inside at ~200 mm length
The reactor where the conversion of glycerol takes place is also fitted with a K-type
thermocouple [ersquo] placed inside the reactor at length of ~250 mm (or at the top of the reactor
when it is packed with a catalyst) enabling the temperature inside the reactor to be measured
by a digital thermometer It is fed with a mixture of (i) preheated DI water and (ii) the aqueous
glycerol solution A K-type thermocouple connected to a thermometer [H] is installed at the
meeting point of these two solutions to monitor the mixture temperature at ~10 cm before the
inlet of the reactor The two solutions are supplied from two separate reservoirs [a] and [b]
and fed into the reactor [D] in a water to feed ratio of 21 which is achieved by adjusting the
flow rate of the HPLC pumps [A] and [B] The area as well as the connections around the top
and bottom of the furnace is covered with glass wool and then wrapped with aluminium foil to
prevent heat loss The details of all equipment used are provided in Table 25
148
Table 25 Details of equipment used in the continuous flow apparatus
Equipment Model or brand
digital thermometer [EH] Comark KM340
electrical ceramic furnace [F] Elite Thermal Systems Limited model TCV12
digital furnace controller [G] Elite Thermal Systems Limited model Eurotherm 2116
HPLC pump [A B] Gilson 305
back pressure regulator [M] Go Regulator model BP-66 max 689 bar
digital pressure gauge [L] Omega max 689 bar
pressure relief valve [K] Ham-let model H900HPSSL14G
II Preparation of feed solution
The feed solution used in this experiment was a combined 06 M aqueous solution of
glycerol and 06 M H2O2 It was prepared by introducing of 5525 g of gt99 wt glycerol (06
mol) and 6803 g of 30 wv H2O2 (06 mol) into a 1000 ml volumetric flask and adjusting the
volume to 1000 ml with DI water
III Preparation of a fixed-bed reactor
When a solid catalyst was used the reactor was loaded with 1 g of a desired zeolite-
based catalyst which was pelletized or extruded prior to administration in order to prevent the
zeolite from being washed out from the reactor The preparation method for pelletized and
extruded catalysts is provided in Chapter 445 and 446 The order of packing of the fixed bed
reactor with the catalyst shown in Figure 37 was arranged as follows The agglomerated
catalyst was placed between two layers of silicon carbide granules (~254 μm) quartz wool and
silica beads (average particle size ~1 mm) acting as an inert filler before being closed off by
149
another layer of quartz wool and metal mesh (200 mesh count per inch) on either ends keeping
the beads in place The temperature gradient inside the reactor shown in Figure 38 was
measured at 125 ordmC at the water flow rate of 243 mlmiddotmin-1 at the reactor depths of 40 30 20
and 10 cm from the top while the temperature of copper block was recorded at a fixed position
ie at 20 cm
Figure 37 The reactor and a K-type thermocouple (left) and the reactor packing scheme (right)
(adapted from Mikkola et al (2014))
150
Figure 38 The temperature gradient inside the reactor
IV Standard operating procedure
The operating procedure of the continuous flow system shown in Figure 35 and 36 is
as follows The HPLC pumps 1 and 2 are turned on pumping the DI water through the system
at a flow rate of 2 mlmin and gradually pressurising the system by closing the back pressure
regulator [M] before heating the furnace up Such order of actions ie the pressurisation at
laboratory temperature followed by increasing of temperature prevented the formation of
steam which is difficult to control The variable back pressure regulator [M] installed on the
system to attain the desired pressure inside the reactor is monitored by a digital pressure gauge
[L] The pressure relief valve [K] is installed to limit the maximum pressure of the system to
200 bar The DI water [a] and feed solution [b] are fed into the preheater [C] and reactor [D]
respectively by high pressure HPLC pumps [A] and [B] at a constant flow rate of 2 mlmin
151
DI water is initially contained in the reservoir [b] before switching to the glycerol feed solution
(a mix of 06 M glycerol and 06 M H2O2) when the desired temperature and pressure are
attained The concentration of glycerol and H2O2 after the mixing point is 02 M
When passing through the heated reactor the feed solution reacts and the fluid then exits
through a coiled tube cooled by an ice-water bath [I] in order to terminate the reactions An in-
line filter [J] made of stainless steel with pore size of 05 μm is used to filter the product solution
before it is passed through the back pressure regulator [M] and a gas-liquid separator [N] When
the samples pass the back pressure regulator they are under atmospheric pressure The liquid
samples are collected at the bottom exit tube [O] while the gaseous products are released at the
top exit through a plastic tube [P] connected to a fume extractor
The samples were collected when the temperature and pressure of the reactor stabilised
at around the set point (plusmn1 degC and plusmn1 bar respectively) and approx 100 ml of the eluted liquid
was collected to ensure consistency To measure the decomposition of H2O2 500 microl of the
liquid samples were immediately mixed with 1000 microl of the titanium sulphate solution and the
UV-VIS spectrophotometry analysis was carried out in 1-5 h The samples for 1H-NMR
analysis were also prepared promptly and analysed 1-5 h later ~5ml of liquid samples were
stored in a fridge for ICP analysis
V Residence time and voidage calculation
The residence times of the feed solution in the reactor were calculated using the water
densities at operating temperature and pressure using Equation 20 and were varied by changing
the flow rate of the pumps
152
Equation 20 Formula calculation of residence time (120591) based on the water densities at
operating temperature and pressure (Bicker et al 2005)
τ =VR 120576 ρR
V ρSTP
Where VR is the reactor volume cm3
120576 is the void fraction of the bed (which is 1 for the empty-tube experiment)
ρR is the fluid density at process pressure and temperature gcm3
ρSTP is fluid density at standard temperature and pressure (0 and 101325Pa) gcm3
V is the volume flow cm3
The fluid density is assumed to be of pure water in this study because of the low concentration
of glycerol
When a reactor is packed with catalyst the reacting fluid may channel through void
fraction Consequently the molecules do not spend as much time in the reactor as those in a
flow through reactor (Fogler 2008)
Voidage ie the bed volume fraction unoccupied by solid material was determined
experimentally by the displacement method using water The packed bed reactor was weighted
connected to an HPLC pump filling the reactor with water and reweighed The weight
difference (4283 g) was then subtracted by the amount of water in the catalyst pores
(0111135cm3g) determined by a BET N2 analysis see Equation 21
153
Equation 21 Determination of packed bed reactor voidage 120576 containing 1g of catalyst
ε =119881119900119897119906119898119890 119900119891 119908119886119905119890119903 119894119899119904119894119889119890 119905ℎ119890 119903119890119886119888119905119900119903 minus 119898119894119888119903119900119901119900119903119890 119907119900119897119906119898119890 119900119891 119905ℎ119890 119888119886119905119886119897119910119904119905
119879119900119905119886119897 119907119900119897119906119898119890 119900119891 119905ℎ119890 119903119890119886119888119905119900119903
Therefore
ε =4283 g divide 0999 gcm3 minus 1g times 0111135 cm3g
74993 cm3= 0556
The fluid density dependence on temperature and pressure may be expressed by Equation 22
Equation 22 Formula for calculation of fluid density at various temperature and pressure
(wwwengineeringtoolboxcom)
ρR =ρSTP
[1 + β(T1 minus T0)] ∙ [1 minus (P1 minus P0)E]
Where
ρR is the fluid density at process pressure and temperature gcm3
ρSTP is the fluid density at Standard Temperature and Pressure (0 and 101325Pa) gcm3
β is the volumetric temperature expansion coefficient m3m3
T1 is the final temperature
T0 is the initial temperature (standard temperature)
P1 is the final pressure Pa
P0 is the initial pressure (standard pressure) Pa
E is the bulk modulus fluid elasticity Pa
154
The water density at different processing pressures and temperatures as well as the total
volume flow rate at different resident times as a function of temperature are provided in Table
56 in the Appendix
155
5 Catalyst characterisation
This chapter presents the results of catalyst characterisation The techniques used and
the experimental procedures are detailed in Chapter 41 and 44 respectively
51 H-Beta zeolite and H-ZSM-5
The successful preparation of H-zeolites by calcination of NH4-zeolites at the
temperatures between 500-550 degC was reported by a number of studies (Kerr 1969 Bagnasco
1996 Kuumlhl 1999 Barthos et al 2000 Lenarda et al 2003) H-Beta zeolite and H-ZSM-5
were prepared by calcining the corresponding commercially available NH4-Beta zeolite
(SiO2Al2O3 = 25) and NH4-ZSM-5 (SiO2Al2O3 = 23) at 550 degC for 5 h in air atmosphere at a
heatingcooling rate of 100 degCh The resulting samples were characterized by XRD XRF
SEM-EDS and TPD-NH3
511 Powder X-ray diffraction
The XRD spectra of NH4-ZSM-5 and NH4-Beta zeolite before and after the calcination
at 550 degC for 5 h are shown in Figure 39 The XRD patterns of the zeolites before and after the
thermal treatment are nearly identical and do not reveal the presence of other distinct phases
which is in accordance with the observation made by other researchers (Trombetta et al 2000
Roberge et al 2002 Lenarda et al 2003 Drake et al 2006) However the intensities of the
main diffraction peaks of the calcined NH4-Beta zeolite (Figure 39C) at around 2θ = 78deg and
2θ = 226deg are ~28 and ~5 higher than those of non-calcined NH4-Beta zeolite (Figure
39D) respectively indicating an increase in crystallinity as a result of thermal treatment An
increase in the crystallinity of the treated samples may possibly be explained by the removal
156
of amorphous phase present in the untreated sample andor reduced distortion of the zeolite
framework (Ravenelle et al 2010) Additionally the SEM micrograph of NH4-Beta zeolite
after calcination shows as opposed to that before calcination individual particles with
possibly-increased crystallinity see Figure 40
An increase in the crystallinity of the treated samples is in contrary to the observation
reported by Lenarda et al (2003) who reported no changes in crystallinity after the calcination
of NH4-Beta zeolite in air for 16 h at 500 600 650 and 750 degC at an un-specified heating rate
The crystallographic data reported do not allow the changes in peak intensities if any to be
confirmed The chemical composition of NH4-Beta zeolite used was identical to that used in
the present work
In general the changes in the peak intensity of the calcined NH4-Beta zeolite may be
induced by (i) the nature of balancing cations andor (ii) the in-channel contents of zeolites
(Medrud 1999) The H+ form of zeolite has a lower X-ray radiation absorption coefficient than
its NH4+ form resulting in higher intensities of diffraction peaks (Hubbell and Seltzer 1995)
The low angle diffraction peaks of zeolites are known to be the most sensitive to the changes
in channel content (Medrud 1999) hence an increase in the intensity of the peak at 2θ = 78deg
is likely to be associated with the decrease of zeolite water content during calcination at 550
degC see Figure 118 (in the Appendix) for the thermogravimetric analysis of NH4-Beta zeolite
The calcination of NH4-ZSM-5 as opposed to NH4-Beta zeolite at 550 degC for 5 h did
not result in a significant change in crystallinity as the intensity of its main characteristic peak
at around 2θ = 2303deg increased only by 1 (see Figure 39A amp B) possibly due to the fact that
the structure of ZSM-5 is more rigid than that of Beta zeolite (Gonzalez et al 2011)
157
10 20 30 40 50 60 70
0
5000
10000
0
5000
10000
10 20 30 40 50 60 70
0
5000
10000
0
5000
10000
+ ++
+
NH4-Beta zeolite as received (D)
++
+
NH4-Beta zeolite after calcination (C)
+
++
NH4-ZSM-5 as received (B)
++
+ NaCl
+ NaCl
2 [degree]
2 [degree]
+
+
NH4-ZSM-5 after calcination (A)
Inte
nsi
ty [
co
un
ts]
Inte
nsi
ty [
co
un
ts]
Figure 39 The XRD spectra of (A) NH4-ZSM-5 after calcination (B) NH4-ZSM-5 as received
(C) NH4-Beta zeolite after calcination and (D) NH4-Beta zeolite as received The calcination
was performed at 550 degC for 5 h at a heating rate of 100 degCh NaCl used as an internal standard
at the zeolite to NaCl mass ratio of 267 was added to the samples prior to analysis Symbol +
= NaCl
158
Figure 40 SEM micrographs of NH4-Beta zeolite before (left) and after (right) calcination at
550 ordmC for 5 h
Table 26 summarises the XRD analysis results obtained on NH4-Beta zeolite and NH4-
ZSM-5 before and after calcination The positions of the main diffraction peaks are also listed
A small but detectable shift of the main diffraction peaks of calcined NH4-ZSM-5 and calcined
NH4-Beta zeolite towards higher and lower angles respectively relative to the parent NH4-
zeolites was noted This may be due to the replacement of NH4+ in the zeolite cavity by a
smaller size H+ (Peter et al 2010 Sethia et al 2015) The un-changed position of NaCl peaks
confirms that the shift is not an instrumental artifact
159
Table 26 The XRD analysis results of NH4-Beta zeolite and NH4-ZSM-5 before and after
calcination at 550 degC for 5 h at a heating rate of 100 degCh
Sample Crystallite sizea
[nm]
XRD relative
crystallinityb []
Position of the main
peak (2θ)c [degree]
NH4-ZSM-5 as received 320 100 230
NH4-ZSM-5 after calcination 388 101 231
NH4-Beta zeolite as received 159 100 226
NH4-Beta zeolite after calcination 151 105 225
a Calculated using Scherrerrsquos equation b The intensities of main diffraction peaks of zeolites
relative to the as received NH4-Beta zeolite or NH4-ZSM-5 The error is in the range of 1-5
c The main diffraction peak of Beta zeolite and ZSM-5 is at 2θ = 226deg and 2303deg respectively
corresponding to the major peak of 302 and 501 crystal surfaces
The diffraction peaks of Beta zeolite see Figure 39 are generally broader than those of
ZSM-5 The phenomenon is related to the smaller average crystallite size of Beta zeolite (~12-
16 nm) as compared to ZSM-5 (~31-39 nm) see Table 26 Other factors known to be
contributing to the peak broadening are the crystal lattice imperfections andor the crystal
strains (Camblor et al 1998)
The XRD spectra of NH4-Beta zeolite as well as NH4-ZSM-5 before calcination as
compared to those of after calcination are nearly identical although small detectable changes
in peak positions and peak intensities were noted The differences however do not allow
distinguishing between the NH4+ and H+ forms The samples were further examined for
changes in their physical surface properties using BET analysis see Chapter 512 The
formation of H-zeolites may be evidenced by the changed surface acidity of the samples which
was measured by TPD-NH3 see Chapter 513
160
512 BET N2 analysis
The specific surface area (SSA) of zeolites used in the present work was determined by
BET N2 analysis The BET N2 adsorption and desorption isotherms of (i) as received and
calcined NH4-ZSM-5 and (ii) as received and calcined NH4-Beta zeolites are shown in Figure
41 and 42 respectively The shapes of the adsorption and desorption isotherms of the samples
resemble the type IV isotherm according to the classification by the IUPAC Commission on
Colloid and Surface Chemistry (Balbuena and Gubbins 1993) A steep increase of adsorption
at lower pressure region of the graphs is followed by a plateau indicating the saturation of the
zeolite surface by a gas monolayer The adsorption then continues to increase at higher relative
pressure region indicating a multilayer adsorption The hysteresis loop ie the deviation of the
desorption isotherm from the adsorption isotherm at higher relative pressures is a phenomenon
typically observed on zeolites and is caused by the capillary condensation of adsorbates in
meso- and micropores (Balbuena and Gubbins 1993) The capillary condensation occurs when
the gases accumulated on the walls of the tiny capillary pores of adsorbent condense When the
pressure is reduced during desorption the molecules do not tend to leave their place so higher
pressure drop is required to pull the adsorbed molecules out of their sites Hence the same
molecules desorb at lower pressure creating a gap between the equilibrium adsorption and
desorption pressures this gap is also known as hysteresis
161
00 01 02 03 04 05 06 07 08 09 10
0
80
100
120
140
160
Qu
an
tity
N2 a
dso
rb
ed
deso
rb
ed
[cm
3g
-1]
Relative Pressure (PPo)
adsorption_calcined NH4-ZSM-5
desorption_calcined NH4-ZSM-5
adsorption_as received NH4-ZSM-5
desorption_as received NH4-ZSM-5
Figure 41 BET N2 adsorptiondesorption isotherms of NH4-ZSM-5 before and after
calcination The data presented is from a single measurement
162
00 01 02 03 04 05 06 07 08 09 10
0
100
200
300
400
500
600
Qu
an
tity
N2 a
dso
rb
ed
deso
rb
ed
[cm
3g
-1]
Relative Pressure (PPo)
adsorption_calcined NH4-Beta zeolite
desorption_calcined NH4-Beta zeolite
adsorption_as received NH4-Beta zeolite
desorption_as received NH4-Beta zeolite
Figure 42 BET N2 adsorptiondesorption isotherms of NH4-Beta zeolite before and after
calcination The data presented is from a single measurement
The surface areas of NH4-Beta zeolite and NH4-ZSM-5 before and after calcination
determined by BET N2 analysis are summarised in Table 27 The BET N2 surface area of NH4-
Beta zeolites after the calcination increased by ~4 while that of calcined NH4-ZSM-5
increased by 21 possibly due to the removal of bulky NH3 allowing higher accessibility of
the internal surface by the N2 molecules The surface area of the material is also related to its
primary crystallite size The Beta zeolites having primary crystallite sizes of ~12-16 nm have
the BET N2 surface area of ~575-600 m2middotg-1 while that of ZSM-5 with the crystallite size of
~31-39 nm is ~300-360 m2middotg-1
163
Table 27 BET N2 surface analysis results of NH4-Beta zeolite and NH4-ZSM-5 before and
after the calcination
Sample BET N2
surface area
[msup2middotg-1]
Micropore
volume
[cmsup3middotg-1]
Micropore
area
[msup2middotg-1]
External
surface area
[msup2middotg-1]
NH4-ZSM-5 as received 299 013 257 42
NH4-ZSM-5 after calcination 363 012 258 105
NH4-Beta zeolite as received 575 017 358 217
NH4-Beta zeolite after calcination 598 017 367 231
513 TPD-NH3
In TPD-NH3 analysis the temperature at which NH3 desorbs correlates with the
strength of the acid sites (see Figure 43 and 44) while the amount of desorbed NH3 reflected
in the area under the concentration of NH3 against time curves is directly proportional to the
total number of acid sites in the catalyst (see Figure 45) In this work the individual desorption
peaks were deconvoluted using Gaussian model and the areas under the peaks were determined
by integration within the time range from 0-6000 s corresponding to the desorption
temperature of 100-600 degC It should be noted that a small but negligible desorption was
detected at le 100 degC thus introducing a negligible experimental error lt1 The absolute
number of zeolite acid sites was determined by multiplying the area under the peak by the total
gas flow rate used
The TPD-NH3 profiles of the calcined NH4-ZSM-5 and calcined NH4-Beta zeolite
compared to the as received NH4-forms are shown in Figure 43 and 44 respectively See Table
28 for the interpretation of NH3 desorption peaks identified in the TPD-NH3 profiles of NH4-
ZSM-5 and NH4-Beta zeolite before and after calcination Figure 45 shows the peaks labelled
164
as α β and γ deconvoluted from the desorption curves plotted as a function of time and used
for integration The solid curves denote the TPD-NH3 original data while the dashed curves
denote the peaks deconvoluted using the OriginPro 86 graphing software The total amount of
desorbed NH3 and the maximum desorption temperatures (Tmax) of the zeolites are summarised
in Table 29
100 200 300 400 500 600
0
100
200
300
0 1200 2400 3600 4800 6000
Co
ncen
tra
tio
n o
f N
H3 i
n g
as
flo
w [
pp
m]
Temperature [oC]
calcined NH4-ZSM-5 (H-ZSM-5)
NH4-ZSM-5
Time [s]
Figure 43 TPD-NH3 profiles of the as received NH4-ZSM-5 and calcined NH4-ZSM-5 ie
H-ZSM-5
165
100 200 300 400 500 600
0
50
100
150
200
250
300
0 1200 2400 3600 4800 6000
Co
ncen
tra
tio
n o
f N
H3 i
n g
as
flo
w [
pp
m]
Temperature [oC]
calcined NH4-Beta zeolite
(H-Beta zeolite)
NH4-Beta zeolite
Time [s]
Figure 44 TPD-NH3 profiles of the as received NH4-Beta zeolite and calcined NH4-Beta
zeolite (or H-Beta zeolite)
Table 28 The interpretation of NH3 desorption peaks identified in the TPD-NH3 profiles of
NH4-ZMS-5 and NH4-Beta zeolite before and after calcination
Desorption peak Temperature range Interpretation
Low temperature peak 1
(LT-peak 1 or α peak)
lt150 degC The desorption peak of physically adsorbed
NH3 (Hegde et al 1989)
Low temperature peak 2
(LT-peak 2 or β peak)
150-260 degC This peak is associated with weakly bonded
NH3 possibly adsorbed on Lewis acid sites
(Topsoslashe et al 1981 Bagnasco 1996
Brandenberger et al 2009) Often part of the
desorbed NH3 is weakly held or physically
adsorbed on sites such as silanol groups and it
is therefore not related to the acid sites (Hegde
et al 1989 Niwa and Katada 1997)
High temperature peak
(HT-peak or γ peak)
260-577 degC The desorption peak of NH3 adsorbed on acid
sites (Niwa and Katada 1997)
166
0 1000 2000 3000 4000 5000 6000
0
50
100
Experimental data
Gaussian deconvolution
Co
nce
ntr
ati
on
of
NH
3 i
n g
as
flo
w [
pp
m]
Time [s]
(A) NH4-ZSM-5
0 1000 2000 3000 4000 5000 6000
0
50
100
150
200
250
300
Experimental data
Gaussian deconvolution
Con
cen
trati
on
of
NH
3 i
n g
as
flow
[p
pm
]
Time [s]
(B) H-ZSM-5
0 1000 2000 3000 4000 5000 6000
0
50
100
150
200
250
300
Experimental data
Gaussian deconvolution
Con
cen
trati
on
of
NH
3 i
n g
as
flow
[p
pm
]
Time [s]
(C) NH4-Beta zeolite
0 1000 2000 3000 4000 5000 6000
0
50
100
150
200
250
Experimental data
Gaussian deconvolution
Co
ncen
tra
tio
n o
f N
H3 i
n g
as
flo
w [
pp
m]
Time [s]
(D) H-Beta zeolite
Figure 45 TPD-NH3 profiles of (A) NH4-ZSM-5 (B) H-ZSM-5 (C) NH4-Beta zeolite and (D) H-Beta zeolite plotted as a function of time The
black solid lines denote the TPD-NH3 original data while the red dotted lines denote the deconvoluted data
167
Table 29 Total amount of desorbed NH3 and the maximum desorption temperatures (Tmax) of NH4-Beta zeolite H-Beta zeolite NH4-ZSM-5 and
H-ZSM-5 The data presented is from a single measurement unless stated otherwise
Catalyst
α-peak β-peak γ-peak NH3
desorbed
[micromolmiddotg-1]
Ref Tmax
[degC]
NH3
[micromolmiddotg-1]
Tmax
[degC]
NH3
[micromolmiddotg-1]
Tmax
[degC]
NH3
[micromolmiddotg-1]
NH4-ZSM-5 145 4 200 69 420 67 140 This work
H-ZSM-5 145 8 187 242 385 269a 520 This work
150 - 388 - - Long and Yang (2001)
NH4-Beta 127 38 162 amp 250 109 amp 298 - - 445 This work
- - - - 630 Alotaibi et al (2014)
H-Beta 124 2 176 105 300 236a 344 This work
277 - 377 - 750 Lenarda et al (2003)
197 653 363 209 862 Talebi et al (2008)
LT-peak = low temperature peak HT= high temperature peak a The error is in the range of 5-10
168
I ZSM-5
The NH3-desorption profile of the as received NH4-ZSM-5 shows at least two
overlapping peaks with the maximums at ~145 and 200 degC (low temperature peak LT-peak)
and another peak at 420 degC (high temperature peak HT-peak) see Figure 43 These
overlapping peaks were deconvoluted into α- β- and γ peaks respectively (see Figure 45A)
Similarly two major desorption peaks are observed for calcined NH4-ZSM-5 an LT-peak at
187 degC and an HT-peak at 385 degC see Figure 43 and 45B The LT-peak also exhibits a shoulder
highly likely due to overlapping of the α and β peaks that otherwise would have appeared at
around 145 and 200 degC as in the TPD curve of the as received NH4-ZSM-5 The extensive
overlapping of the α and β peaks is usually observed in TPD of NH3 adsorbed at low
temperature ie at 20-30 degC (Topsoslashe et al 1981 Hegde et al 1989 Hunger et al 1990
Brandenberger et al 2009) hence the overlapping was not expected in the present work as the
NH3 adsorption step was performed at 100 degC The phenomenon may be explained by the
insufficient time used ie 1 h for the removal of excess of NH3 before desorption
Different NH3 desorption peaks are attributed to acidic centres of different strengths
According to Brandenberger et al (2009) who studied the surface acidity of H-ZSM-5 (SiAl
~10) and Fe-exchanged ZSM-5 using TPD-NH3 and FTIR spectroscopy the LT-peaks at 140
and 230 degC are associated with weakly adsorbed NH3 on Lewis acid sites while the HT-peak
at 490 degC is related to NH4+ ions with three hydrogen atoms bonded to three oxygen ions of
AlO4 tetrahedra (3H structure) on Broslashnsted acid sites see Figure 46 for a schematic illustration
of NH3 adsorbed on different zeolite sites
169
Figure 46 Schematic illustration of NH3 adsorbed on different Broslashnsted acidic centres (adapted
from Rodriacuteguez-Gonzaacutelez et al (2008))
The above interpretation is in agreement with that previously reported by Topsoslashe et al
(1981) Bagnasco (1996) and Rodriacuteguez-Gonzaacutelez et al (2008) Similarly Niwa and Katada
(1997) identified the HT-peak (277-577 degC) as the desorption peak of NH3 adsorbed on the
acid sites but the LT-peak (127-277 degC) was assigned to NH3 weakly held or physically
170
adsorbed on the zeolite and it is therefore not ascribable to the acid sites See Table 28 for a
summary of NH3 desorption peaks identified in the present work
The presence of LT-peaks in the NH3-desorption profile of the as received NH4-ZSM-
5 is observed in this work but it was not reported by Long and Yang (2001) possibly due to
different removal degrees of physically adsorbed NH3 as a result of different zeolite weight to
flow rate of carrier gas ratio (WF) (Niwa and Katada 1997) In fact a direct comparison is
impossible since the peak shapes and positions not only depend on experimental conditions
such as the sample weight gas flow and heating rate but also on the mechanisms controlling
the desorption such as the re-adsorption andor diffusion (Niwa and Katada 1997 Rodriacuteguez-
Gonzaacutelez et al 2008) The appearance of the HT-peak may be related to the formation of H-
form of the zeolite during the sample degassing in He flow at 550 degC for 1 h
Due to the extensive-overlapping of the two LT-peaks (α and β peaks) as well as the
uncertainty about the interpretation of the β-peak (Hunger et al 1990 Bagnasco 1996 Niwa
and Katada 1997 Rodriacuteguez-Gonzaacutelez et al 2008 Brandenberger et al 2009) only the HT-
peak is used for quantitative determination of the zeolite acid sites Hence the concentration
of the Broslashnsted acid sites of NH4-ZSM-5 before and after the calcination determined was 67
and 269 micromolmiddotg-1 respectively This may be considered as the key evidence indicating that the
as received NH4-ZSM-5 was successfully transformed into its protonic form ie H-ZSM-5
II Beta zeolite
The desorption profile of the as received NH4-Beta zeolite see Figure 44 shows only
three overlapping LT-peaks ie α- β1- and β2- peaks at 128 165 and 250 degC respectively see
Figure 45C for the deconvolution of the overlapped peaks These peaks are highly likely to be
171
associated with NH3 adsorbed on structural defect sites present in significant quantity in Beta
zeolite (Camiloti et al 1999) The high intensities of the α- and β1- peaks indicate that part of
the desorbed NH3 may also be weakly held or physically adsorbed on zeolite sites such as
silanol groups (Hegde et al 1989) According to Niwa and Katada (1997) the concentration
of weak acid sites corresponding to the β-peak was found to be of no catalytic relevance
explaining the low conversion of glycerol observed in the current work see Figure 83 (page
258) The interpretation of the NH3 desorption peaks is summarised in Table 28
Two broad peaks including a β-peak with a small shoulder at 176 degC and a γ-peak at
300 degC are observed in the NH3 desorption curve of the calcined NH4-Beta zeolite (see Figure
44 and 45D) as opposed to the observation of Lenarda et al (2003) who reported two
desorption peaks at ~277 degC and ~377 degC The difference in the peak position may be related
to the WF ratio andor the re-adsorption of NH3 (Niwa and Katada 1997)
Similar to the case of ZSM-5 the extensive-overlapping of the two LT-peaks (α and β
peaks) and the different interpretation of the β peak means that only the HT-peak can be used
in calculating the concentration of acid sites While the concentration of acid sites of the as
received NH4-Beta zeolite could not be determined that of the calcined NH4-Beta zeolite was
~233 micromolmiddotg-1 indicating successful transformation of NH4-Beta zeolite to H-Beta zeolite
Thus the calcined NH4-Beta zeolite is referred to as H-Beta zeolite hereafter in this report
The TPD-NH3 profiles of Beta zeolite and ZSM-5 are similar to those reported in
literature however a different desorption rate and the maximum desorption temperatures (Tmax)
were reported (Long and Yang 2001 Lenarda et al 2003 Talebi et al 2008) In general the
extent of NH3 desorption determined on both NH4-Beta zeolite and NH4-ZMS-5 before and
after calcination was found lower than previously reported (Lenarda et al 2003 Talebi et al
172
2008) possibly due to different experimental conditions and the peak assignation to acid sites
It should be noted that the quantification of NH3 in the present work was based on the peak of
me=17 in the mass spectra to which the fragmented water is known to contribute (Akah et al
2005) Hence the intensity of NH3 in mass spectra obtained was reduced by 20 to account
for the contribution of water
52 Metal-exchanged zeolites
The preparation of transition metal-exchanged zeolites by an ion exchange reaction is
typically carried out using the zeolites in their H-form as the preferred starting materials as
opposed to the NH4-form known to exhibit lower catalytic potential (Naccache and Taarit
1984 Kucherov and Slinkin 1998 Baek et al 2004 Marakatti et al 2014) However several
recent studies reported using the NH4-form of zeolites in the preparation of metal-exchanged
zeolites achieving a high degree of ion exchange (Mihaacutelyi et al 1997 Li and Armor 1999
Mavrodinova et al 2001 Habib et al 2014) The possibility of avoiding an energy intensive
calcination step associated with preparation of H-form of zeolite is attractive To investigate
whether the use of H- and NH4-forms in the preparation of metal-exchanged zeolites would
result in final products with identical or competitive properties both forms of Beta zeolite and
ZSM-5 were used in the present work The 48 wt metal-exchanged Beta zeolite and ZSM-5
catalysts were prepared by solid state ion exchange (SSIE) using various metal salt precursors
including Ce(NO3)3middot6H2O La(NO3)3middotH2O Sn(CH3COO)2 and Zn(CH3COO)2 according to the
experimental protocol described in Chapter 442 The resulting samples were characterized by
XRD XRF TEM-EDS and TPD-NH3 The preparation of metal-exchanged zeolites containing
48 wt of metal relative to zeolite in the initial mixture before the calcination was chosen
173
based on the work of De La Torre et al (2012) Due to the different molar weights of used
metals only a partial ion exchange is expected
521 Powder X-ray diffraction
The XRD spectra of (i) H-Beta zeolite and (ii) H-ZSM-5 before and after SSIE reaction
with various metal salts ie 48 wt MH-Beta zeolite and 48 wt MH-ZSM-5 where M =
Ce La Sn and Zn are shown correspondingly in Figure 47 and 48 The XRD spectra of NH4-
Beta zeolite NH4-ZSM-5 and their MNH4-forms are provided in the Appendix in Figure 119
and 120 respectively
The XRD patterns of 48 wt MH-zeolites (Figure 47 and 48) and 48 wt MNH4-
zeolites (Figure 119 and 120) are nearly identical except for the slightly higher intensities of
the main diffraction peaks of MH-zeolites (2θ = 226deg for Beta zeolite 2θ = 2303deg for ZSM-
5) as compared to MNH4-zeolites The XRD patterns of all metal-exchanged zeolites studied
except for 48 wt Lazeolites show the traces of corresponding metal oxides ie ZnO SnO2
and CeO2 La2O3 was detected neither in 48 wt LaBeta zeolite nor 48 wt LaZSM-5 This
may be possibly explained by incomplete thermal decomposition of La(NO3)3middotH2O to
La3O4NO3 at 550 degC (Mentus et al 2007) No other phases were detected
The intensities of the diffraction peaks of the parent zeolites ie NH4-Beta zeolite H-
Beta zeolite NH4-ZSM-5 and H-ZSM-5 show a tendency to decrease considerably after the
SSIE reaction with metal salts For instance the intensity of the main diffraction peak of 48
wt CeNH4-Beta zeolite decreased to 55 as compared to NH4-Beta zeolite see Table 30
Such phenomenon is usually observed when metal ions enter the zeolites structure due to
(Medrud 1999 Rauscher et al 1999)
174
i their higher absorption coefficient of the X-ray radiation and
ii the lower relative zeolite content in the samples due to the addition of the metal salts
In addition to a decrease in the intensity the main diffraction peaks of the parent zeolites
(2θ = 226deg for Beta zeolite 2θ = 2303deg for ZSM-5) are observed to shift after the SSIE This
may be due to the replacement of H+ or NH4+ in the zeolite cavity by a different size ion (Peter
et al 2010 Sethia et al 2015) The shift does not show a consistent trend (see Table 30) In
general the SSIE of the zeolites caused their main diffraction peaks to shift to higher angles
for Beta zeolite but towards lower angles for ZSM-5 with respect to their H-forms Although
the un-changed position of NaCl peaks indicates that the shift of the main diffraction peaks of
the zeolite samples is unlikely to be an instrumental artifact the extents of the shifts are
considered as insignificant due to being about the same as the 2θ angle step size used for the
XRD measurement ie ~0033deg The positions of the main diffraction peaks crystallite sizes
and the relative crystallinities of Beta zeolite and ZSM-5 after SSIE compared to their
corresponding parent NH4- and H-forms are summarised in Table 30
175
10 20 30 40 50 60 70
0
5000
10000
0
5000
10000
0
5000
10000
0
5000
10000
0
5000
10000
+x
x
+
+
xx
CeO2
48 wt CeH-Beta
+++
+
48 wt LaH-Beta
O
++
OOO
48 wt SnH-Beta
ZnO
O
++
+
+
x
SnO2
48 wt ZnH-Beta
+
+
+
Inte
nsi
ty [
co
un
ts]
2 [degree]
+ NaCl
H-Beta
+
Figure 47 The XRD patterns of H-Beta zeolite and 48 wt MH-Beta zeolite where M = Ce
La Sn and Zn prepared by SSIE using corresponding metal salt as a precursor at 550 degC for
3 h at a heating rate of 100 degCh NaCl used as an internal standard at the zeolite to NaCl mass
ratio of 267 was added to the samples prior to analysis Symbol + = NaCl ZnO ο = SnO2
x = CeO2
176
10 20 30 40 50 60 70
0
5000
10000
0
5000
10000
0
5000
10000
0
5000
10000
0
5000
10000
+X+
+
+
X
X
2 (degree)
48 wt CeH-ZSM5
X
+ ++
+
OOO
48 wt LaH-ZSM5
+
+
O
48 wt SnH-ZSM5
O
+++
+
48 wt ZnH-ZSM5
Inte
nsi
ty [
cou
nts
]
+ NaCl ZnO O SnO2
x CeO2
+ +
+
H-ZSM5
+
Figure 48 The XRD patterns of H-ZSM-5 and 48 wt MH-ZSM-5 where M = Ce La Sn
and Zn prepared by SSIE using corresponding metal salt as a precursor at 550 degC for 3 h at a
heating rate of 100 degCh NaCl used as an internal standard at the zeolite to NaCl mass ratio of
267 was added to the samples prior to analysis Symbol + = NaCl ZnO ο = SnO2 x = CeO2
177
Table 30 The positions of the main diffraction peaks the crystallite sizes and the relative
crystallinities of NH4-Beta zeolite H-Beta zeolite NH4-ZSM-5 H-ZSM-5 before and after the
SSIE using various metals salts as precursors The percentages reported are in wt indicating
the metal content relative to zeolite
Sample Crystallite
sizea [nm]
XRD relative
crystallinityb
[]
Position of the
main peakc
(2θ) [degree]
Shift in 2θ of
the main peakc
[degree]
NH4-Beta 159 100 226 +009
H-Beta 151 105 225 000
48 CeNH4-Beta 145 55 225 +007
48 LaNH4-Beta 141 64 225 +003
48 SnNH4-Beta 144 47 225 +002
48 ZnNH4-Beta 121 76 225 -001
48 CeH-Beta 146 64 225 +007
48 LaH-Beta 137 68 226 +008
48 SnH-Beta 139 63 225 +004
48 ZnH-Beta 135 83 225 +006
NH4-ZSM-5 320 100 230 -005
H-ZSM-5 388 101 231 000
48 CeNH4-ZSM-5 331 60 230 -005
48 LaNH4-ZSM-5 372 71 231 -003
48 SnNH4-ZSM-5 381 65 231 000
48 ZnNH4-ZSM-5 369 98 231 000
48 CeH-ZSM-5 318 62 231 -003
48 LaH-ZSM-5 345 75 231 -003
48 SnH-ZSM-5 370 69 231 +003
48 ZnH-ZSM-5 351 84 231 -003
a Calculated using Scherrerrsquos equation b The intensities of main diffraction peaks of zeolites
relative to the NH4-Beta zeolite or NH4-ZSM-5 The error is in the range of 1-7 c The main
diffraction peak of Beta zeolite and ZSM-5 is at 2θ = 226deg and 2303deg respectively
corresponding to the peak of 302 and 501 crystal surfaces
178
522 BET N2 analysis
Figure 49 shows the BET N2 adsorptiondesorption isotherms of (A) H-Beta zeolite and
48 wt SnH-Beta zeolite and (B) H-ZSM-5 and 48 wt SnH-ZSM-5 zeolite respectively
These isotherms are the representatives of BET N2 adsorptiondesorption isotherms of H-Beta
zeolite and H-ZSM-5 before and after the SSIE reaction with various metal salts including
Ce(NO3)3middot6H2O La(NO3)3middotH2O and Zn(CH3COO)2 Similar to H-Beta zeolite and H-ZSM-5
(see Chapter 512) the BET N2 adsorptiondesorption isotherms of 48 wt Mzeolites
prepared in the present work resemble the type IV adsorption isotherm according to the
classification by the IUPAC Commission on Colloid and Surface Chemistry (Balbuena and
Gubbins 1993) See Chapter 512 for a description of the type IV adsorption isotherm
00 01 02 03 04 05 06 07 08 09 10
0
100
200
300
400
500
600
00 01 02 03 04 05 06 07 08 09 10
0
80
100
120
140
160
Qu
an
tity
N2 a
dso
rb
ed
deso
rb
ed
[cm
3g
-1]
Relative Pressure (PPo)
adsorption_H-Beta zeolite
desorption_H-Beta zeolite
adsorption_48 SnH-Beta zeolite
desorption_48 SnH-Beta zeolite
(A) (B)
Qu
an
tity
N2 a
dso
rb
ed
deso
rb
ed
[cm
3g
-1]
Relative Pressure (PPo)
adsorption_H-ZSM-5
desorption_H-ZSM-5
adsorption_48 SnH-ZSM-5
desorption_48 SnH-ZSM-5
Figure 49 N2 adsorptiondesorption isotherms of (A) H-Beta zeolite and 48 wt SnH-Beta zeolite
and (B) H-ZSM-5 and 48 wt SnH-ZSM-5 The data presented is from a single measurement
The BET analysis results of all the 48 wt Mzeolites prepared are summarised in
Table 31 The measured BET N2 surface areas of NH4-Beta zeolite H-Beta zeolite NH4-ZSM-
179
5 and H-ZSM-5 considerably decreased after the SSIE possibly due to the blockage of the
zeolite pores by metal aggregates De La Torre et al (2012) also observed a decrease in BET
N2 surface area after incorporation of copper to H-Beta zeolite (SiAl ~25) and H-ZSM-5 (SiAl
~50) by liquid ion exchange and wet impregnation methods
Table 31 BET N2 surface analysis results of NH4-Beta zeolite H-Beta zeolite NH4-ZSM-5
and H-ZSM-5 before and after SSIE reaction with various metal salts The data presented is
from a single measurement
Sample
BET N2
surface area
[msup2g-1]
Micropore
volume
[cmsup3g-1]
Micropore
area
[msup2g-1]
External
surface area
[msup2g-1]
NH4-Beta 575 017 358 218
H-Beta 598 017 367 230
48 CeNH4-Beta 534 015 330 204
48 LaNH4-Beta 591 015 383 208
48 SnNH4-Beta 489 014 303 186
48 ZnNH4-Beta 526 015 326 200
48 CeH-Beta 539 015 333 206
48 LaH-Beta 567 015 371 196
48 SnH-Beta 514 014 301 213
48 ZnH-Beta 528 014 313 215
NH4-ZSM-5 299 013 257 42
H-ZSM-5 363 012 258 104
48 CeNH4-ZSM-5 294 010 213 81
48 LaNH4-ZSM-5 370 011 280 90
48 SnNH4-ZSM-5 323 011 228 95
48 ZnNH4-ZSM-5 336 012 250 86
48 CeH-ZSM-5 340 011 235 104
48 LaH-ZSM-5 359 011 284 76
48 SnH-ZSM-5 304 010 223 80
48 ZnH-ZSM-5 308 011 227 81
180
523 TPD-NH3
The transformation of Broslashnsted acid sites of zeolites into Lewis acid sites was attempted
by an ion-exchange reaction replacing the balancing H+ with metal ions including Ce La Sn
and Zn The degree of ion exchange was quantified by the decrease in concentration of
Broslashnsted acid sites determined by TPD-NH3
The TPD-NH3 profiles of (A) NH4-Beta zeolite (B) H-Beta zeolite (C) NH4-ZSM-5 and
(D) H-ZSM-5 before and after the SSIE reaction with various metal salts are shown in Figure
50 In general the TPD-NH3 profiles of all metal-exchanged zeolites prepared are similar to
the profiles of their parent NH4- or H-zeolites (see Page 164-165 ) two main desorption peaks
including a low temperature peak (LT-peak) and a high temperature peak (HT-peak)
corresponding to β- and γ-peak respectively are observed See Chapter 513 for the
interpretation of NH3 desorption peaks The LT-peak seems to overlap another smaller peak
appearing as a shoulder at the Tmax = 124-149 degC highly likely resulting from an extensive-
overlapping of the β-peak and another LT-peak referred to as an α-peak appearing at the
temperatures lt150 degC
The TPD-NH3 profiles of all the metal-exchanged zeolites studied exhibited
(i) a shift of the maximum desorption temperatures (Tmax) toward higher
temperatures 1-17 ordmC shift was recorded for all samples except for 48 wt Ce-
La- and ZnH-ZSM-5 which exhibited a shift towards lower temperatures by
about 5 degC
(ii) a lower amount of desorbed NH3
The amount of NH3 desorbed associated with the γ peak and the Tmax of individual peaks in the
TPD-NH3 profiles recorded for all the zeolites tested are summarised in Table 32
181
Figure 50 TPD-NH3 profiles of (A) NH4-Beta zeolite (B) H-Beta zeolite (C) NH4-ZSM-5 and (D) H-ZSM-5 before and after SSIE reaction with
various metal salts at 550 degC 3 h at a heating rate of 100 degCh
100 200 300 400 500 600
0
50
100
150
200
250
300
100 200 300 400 500 600
0
50
100
150
200
250
300
100 200 300 400 500 600
0
50
100
150
200
250
300
100 200 300 400 500 600
0
50
100
150
200
250
300
48 CeNH4-Beta zeolite
48 LaNH4-Beta zeolite
48 ZnNH4-Beta zeolite
Co
ncen
tra
tio
n o
f N
H3
in
ga
s fl
ow
[p
pm
]
Temperature [oC]
(A)
NH4-Beta zeolite
48 SnNH4-Beta zeolite
(C)
Co
ncen
tra
tio
n o
f N
H3
in
ga
s fl
ow
[p
pm
]
Temperature [oC]
NH4-ZSM-5
48 SnNH4-ZSM-5
48 ZnNH4-ZSM-5
48 LaNH4-ZSM-5
48 CeNH4-ZSM-5
(B)
Co
ncen
tra
tio
n o
f N
H3
in
ga
s fl
ow
[p
pm
]
Temperature [oC]
H-Beta zeolite
48 SnH-Beta zeolite
48 ZnH-Beta zeolite
48 LaH-Beta zeolite
48 CeH-Beta zeolite
(D)
Co
ncen
tra
tio
n o
f N
H3
in
ga
s fl
ow
[p
pm
]Temperature [
oC]
48 CeH-ZSM-5
48 LaH-ZSM-5
48 ZnH-ZSM-5
H-ZSM-5
48 SnH-ZSM-5
182
Table 32 The amount of desorbed NH3 associated with the γ peak and the maximum desorption
temperatures (Tmax) of individual peaks in TPD-NH3 profiles of NH4-Beta zeolite H-Beta
zeolite NH4-ZSM-5 and H-ZSM-5 before and after SSIE with various metal salts The metal
contents are given in wt relative to zeolite
The data presented is from a single measurement unless stated otherwise a The error is in the
range of 5-10 b Determined using the following formula D = (a-b)c where a = the number of
strong Broslashnsted acid site (γ-peak) present in H-Beta zeolite or H-ZSM-5 b = the number of strong
Broslashnsted acid site (γ-peak) in metal-exchanged zeolite c = the number of formal charges of the
incoming metal cation
Catalyst
α-peak β-peak γ-peak
Degree of ion
exchange b Tmax
[degC]
Tmax
[degC]
Tmax
[degC]
NH3
[micromolmiddotg-1]
NH4-Beta 127 165250 - - 0
H-Beta 124 176 300 236 a 0
48 CeNH4-Beta 134 177 306 94 36
48 LaNH4-Beta 138 178 317 107 43
48 SnNH4-Beta 136 171 304 147 22
48 ZnNH4-Beta 148 178 305 114 61
48 CeH-Beta 140 160 305 73 41
48 LaH-Beta 145 176 317 95 47
48 SnH-Beta 141 160 305 45 48
48 ZnH-Beta 149 180 305 103 67
NH4-ZSM-5 145 200 420 67 0
H-ZSM-5 145 187 385 270 a 0
48 CeNH4-ZSM-5 138 190 387 146 31
48 LaNH4-ZSM-5 146 191 400 113 52
48 SnNH4-ZSM-5 133 190 392 39 58
48 ZnNH4-ZSM-5 148 205 386 88 91
48 CeH-ZSM-5 - 187 380 73 49
48 LaH-ZSM-5 144 180 380 98 57
48 SnH-ZSM-5 144 172 391 35 59
48 ZnH-ZSM-5 145 187 380 57 106
183
All the metal-exchanged zeolites analysed by TPD-NH3 show a general trend in a shift
of Tmax towards slightly higher temperatures as compared to H-Beta zeolite and H-ZMS-5 see
Table 32 Due to the extensive-overlapping of the α- and β-peaks as well as the uncertainty
about the origin of desorbed NH3 associated with the β-peak (Hunger et al 1990 Bagnasco
1996 Niwa and Katada 1997 Rodriacuteguez-Gonzaacutelez et al 2008 Brandenberger et al 2009)
only the γ-peak is used for quantitative determination of the zeolite acid sites in this work
Thus only the changes in the γ-peak are discussed herein
The γ-peaks of almost all metal-exchanged NH4- and metal-exchanged H-zeolites shift
towards higher temperatures by 1-17 degC with respect to H-Beta zeolite (Tmax γ = 300 degC) or H-
ZSM-5 (Tmax γ = 385 degC) indicating stronger bonding interaction of NH3 with the Broslashnsted
acid sites This is in contrast to 48 wt Ce- La- and ZnH-ZSM-5 showing an opposite
general trend in the shift of γ-peaks ie a shift by 5 degC towards lower temperature with respect
to the H-ZSM-5
The amount of desorbed NH3 associated with the γ peak is an indicator of the zeolite
surface acidity As expected the surface acidities of the metal-exchanged Beta zeolites are
generally about 50-60 lower than that of H-Beta zeolite except for 48 CeH-Beta zeolite
and 48 SnH-Beta the acidities of which are 70 and 80 lower respectively A significant
decrease in acidity (~60-90) is observed for metal-exchanged ZSM-5 samples In addition to
the ion exchange reaction the decreased surface acidity of the metal-exchanged zeolites is
likely to be associated with the blockage of the zeolite pores by metal aggregates (De La Torre
et al 2012)
The degree of ion exchange was determined by the difference in the number of strong
Broslashnsted acid site (γ-peak) present in the corresponding zeolites before and after ion exchange
divided by the formal charges of the incoming cation (Lugstein et al 1995 Mavrodinova
184
1998) The formal charges of metal cations indicate the number of protons (from SiOHAl
group) required for exchanging with one metal cation Since both the NH4-Beta zeolite and
NH4-ZSM-5 are assumed to transform into their H-forms during the SSIE the term ldquozeolites
before ion exchangerdquo therefore refers to either H-Beta zeolite or H-ZSM-5 The TPD-NH3 data
indicate that the degree of ion exchange between zeolites and metal salts is found to be higher
(2-118 ) when starting from H-zeolites see Table 32 possibly because the NH4-sites get only
partially exchanged and the sites that are not exchanged get deammoniated forming acidic OH
groups (Onyestyak et al 1991 Mavrodinova et al 2001)
The results indicate a higher degree of ion exchange in ZSM-5 (SiAl = 115) samples
than Beta zeolite (SiAl = 125) De La Torre et al (2012) prepared Cu-exchanged Beta zeolite
(SiAl = 25) and ZSM-5 (SiAl = 50) catalysts by liquid ion exchange and found that slightly
higher degrees of ion exchange ranging from 7-25 were achieved in Beta zeolite The
authors speculated that this may be related to the lower SiAl ratios and consequently to the
higher number of sites available for Cu In this work the SiAl ratio of ZSM-5 is slightly lower
than that of Beta zeolite possibly explaining a higher degree of ion exchange in ZSM-5 than
in Beta zeolite
53 Dealuminated H-ZSM-5 and H-Beta zeolite
H-ZSM-5 and H-Beta zeolite were treated with concentrated HNO3 at 80 degC in order to
remove Al atoms from the zeolite frameworks according to the procedure described in Chapter
443 The resulting materials were analysed by XRD XRF and TPD-NH3 The summary of
experimental conditions used for dealumination of H-ZSM-5 and H-Beta zeolite is provided in
Table 33
185
Table 33 The summary of experimental conditions used for dealumination of H-ZSM-5 and
H-Beta zeolite in 20 ml of 13 M HNO3 solution at 100 degC
Starting zeolite Sample label Amount of zeolite used
[g]
Treatment time
[h]
H-ZSM-5 deAl-ZSM-5 1 20
H-Beta deAl-BetaI4hd 1 4
deAl-BetaI6h 1 6
deAl-BetaI20h 1 20
deAl-BetaII4h 2 4
deAl-BetaII6h 2 6
deAl-BetaII20h 2 20
deAl-BetaIII4h 3 4
deAl-BetaII 5min 2 008
deAl-BetaII 15min 2 025
531 Dealuminated ZSM-5
Figure 51 shows the XRD patterns of H-ZSM-5 before (top) and after (bottom) the
treatment with 13 M HNO3 at 100 degC for 20 h The spectra do not reveal significant differences
between before and after the treatment indicating that the MFI structure of ZSM-5 remained
preserved
186
0
300
600
900
10 20 30 40 50 60 70
0
300
600
900
H-ZSM-5In
ten
sity
[C
ou
nts
]
acid treated H-ZSM-5
2 [degree]
Figure 51 XRD spectra of H-ZSM-5 before (top) and after (bottom) the treatment with 13 M
HNO3 at 100 degC for 20 h 1 g of zeolite in 20 ml of 13 M HNO3 solution The angular scan was
between 2θ = 5-70deg with a step size of 0017deg 2θ and a count time of 258 s per step
The morphology and the elemental composition of the acid treated samples was
analysed by the TEM-EDS The results showed that the SiAl ratio increased from 115 to ~122
as a result of the acid treatment indicating a relatively small extent (~6 wt) of dealumination
This is in agreement with the study by Muumlller et al (2000) and Gonzalez et al (2011) The
dealumination of ZSM-5 was reported to be challenging using aqueous HBr H2SO4 or HCl
(Kooyman et al 1997) The low extent of dealumination in ZSM-5 means that the replacement
187
of Al atoms inside the zeolite framework would not be possible Based on the results obtained
further work was abandoned
532 Dealuminated Beta zeolite
5321 Physical properties
H-Beta zeolite was treated with 13 M HNO3 solution at 100 degC for 4 6 and 20 h at the
ratios of 1 2 and 3 g of zeolite in 20 ml of 13 M HNO3 Typical XRD patterns of H-Beta zeolite
before (top) and after (bottom four) the acid treatment for different times are shown in Figure
52 while the related crystallite sizes relative crystallinities as well as the BET-N2 surface areas
and XRF analysis results are summarised in Table 34
188
0
4000
8000
0
4000
8000
0
4000
8000
10 20 30 40 50 60 70
0
4000
8000
+ +
+
+ NaCl H-Beta zeolite
+
+ ++
+
deAl-BetaII4h+ NaCl
Inte
nsi
ty [
Co
un
ts]
+ +
+
+deAl-BetaII6h+ NaCl
+ +
+
+
deAl-BetaII20h+ NaCl
2 [degree]
Figure 52 X-ray diffraction patterns of H-Beta zeolite before and after the treatment with 13
M HNO3 at 100 degC for 4 6 and 20 h deAl-Beta refers to the dealuminated Beta zeolite the
Roman and Arabic numerals indicate the number of grams of H-Beta zeolite dealuminated in
20 ml of 13 M HNO3 and the dealumination time in hours respectively NaCl used as an
internal standard at the zeolite to NaCl mass ratio of 267 was added to the samples prior to
analysis Symbol + = NaCl
189
Table 34 The summary of BET XRD and XRF analysis results of dealuminated Beta zeolites (deAl-Beta) compared to NH4-Beta and H-Beta
zeolites
Catalyst BET-N2
surface area
[msup2middotg-1]
Crystallite
sizea
[nm]
Relative
crystallinityb
[]
2θ of d302
[degree]
shift of d302
[degree]
d302-spacing
[Aring]
SiAl
(XRF)
Al extractedc
[wt]
NH4-Beta zeolite 575 159 100 226 394 124 0
H-Beta zeolite 598 151 104 225 000 396 124 0
deAl-BetaI4hd 573 133 95 227 +017e 391 533 974
deAl-BetaI6h 572 137 94 227 +024 391 600 978
deAl-BetaI20h 554 141 89 226 +007 394 795 983
deAl-BetaII4h 580 129 100 227 +017 393 688 977
deAl-BetaII6h 551 130 94 227 +017 389 671 980
deAl-BetaII20h 542 154 89 226 +009 392 526 969
deAl-BetaIII4h 585 122 95 227 +017 392 556 974
a The primary crystallite size was calculated by Scherrerrsquos equation b Relative crystallinity was calculated using the intensity of the main diffraction
peak (d302 at 2θ = 226deg) considering the NH4-Beta zeolite as a benchmark ie 100 crystalline c Al extracted (wt) was calculated based on the
amount of Al remaining after the acid treatment compared to the initial amount of Al d deAl-Beta refers to dealuminated Beta zeolite the Roman
and Arabic numerals indicate the number of grams of H-Beta zeolite dealuminated in 20 ml of 13 M HNO3 and the dealumination time in hours
respectively e the ldquo+rdquo (plus symbol) denotes the positive peak shift relative to H-Beta zeolite
190
The XRD pattern of H-Beta zeolite after the treatment with HNO3 (see Figure 52) may
be attributed only to Beta zeolite structure (BEA) indicating that the dealumination did not
significantly affect the BEA structure of the Beta zeolite No other distinct phases were
detected However the intensity of the diffraction peak at (i) 2θ = 78deg (d101) increased after 4
and 6 h (deAl-BetaII4h and deAl-BetaII6h) and then considerably decreased after 20 h (deAl-
BetaII for 20 h) while that of the peak at (ii) around 2θ = 226deg slightly decreased but also
substantially shifted to higher angles (see Figure 53) According to Medrud (1999) the
substitution in the tetrahedral framework may cause a substantial shift in 2θ For instance the
substitution of Al3+ for Si4+ causes the unit cell to expand because the Al3+ is a larger ion than
Si4+ Consequently the diffraction peaks tend to shift to lower angles (Camblor et al 1998
Medrud 1999) This implies that any changes in the tetrahedral frameworks would cause a
shift in the 2θ hence the shift of the d302 peak of deAl-Beta zeolites to higher angle in this work
may be ascribable to the removal of Al3+ from the BEA framework A shift of the d302 peak of
Beta zeolite to higher angle after dealumination with 13 M HNO3 solution was also reported
by Dzwigaj et al (2013)
The change in the position of the main diffraction peak of zeolite samples may be used
to obtain information on the lattice contractionexpansion of the framework since according to
the Braggrsquos law equation (see Equation 9 in Chapter 413) the d302 spacing is calculated using
the corresponding 2θ value (Dzwigaj et al 2013) Under the conditions studied the d302
spacing of H-Beta zeolite decreases after the dealumination Independently to the zeoliteHNO3
ratio the d302 spacing of deAl-Beta zeolites decreased after the dealumination for 4 and 6 h
however a slightly bigger d302 spacing was noticed after 20 h dealumination possibly due to
the re-insertion of Si species (Ho et al 1998 Kunkeler et al 1998) An increasedecrease in
the d-spacing in a given series of samples is generally considered a sufficient evidence of lattice
contractionexpansion in the BEA framework It is commonly used for monitoring of the
191
dealumination progress of Beta zeolite as well as the incorporation of metal ions such as
titanium (Ti) silver (Ag) and vanadium (V) ions into the BEA tetrahedral framework (Camblor
et al 1993 Sudhakar Reddy and Sayari 1995 Dzwigaj et al 1998 Dzwigaj et al 2000
Baran et al 2012 Dzwigaj et al 2013) Table 34 summarises the changes of position of the
main diffraction peaks d-spacing (d302) crystallite size and crystallinity of all deAl-Beta
samples as well as their SiAl ratios obtained from XRF analysis and specific surface area from
BET N2 analysis
The change in the intensity of the main diffraction peaks of H-Beta zeolite treated with
HNO3 for different times indicates a possible change in the relative crystallinity see Figure 53
Considering the crystallinity of NH4-Beta zeolite as a benchmark the relative crystallinity of
deAl-Beta samples decreases with the dealumination time The relative crystallinity was found
to be inversely proportional to the crystallite size of deAl-Beta samples On the other hand no
apparent trend was found between the zeolite to HNO3 solution ratio the crystallinity and the
extent of dealumination According to the XRF analysis 97-98 wt of total Al was extracted
from the zeolite framework after 4 h in all samples and remained unchanged for 20 h
confirming the successful removal of Al atoms from the BEA framework see Figure 54
192
6 8 10 20 22 24
0
2000
4000
6000
8000
H-Beta zeolite
deAl-BetaII4h
deAl-BetaII6h
deAl-BetaII20h
Inte
nsi
ty [
cou
nts
]
2 [degree]
Figure 53 The details of the main X-ray diffraction peaks of H-Beta zeolite before and after
dealumination of 2 g of H-Beta zeolite in 20 ml of 13 M HNO3 solution for 4 6 and 20 h The
error is in the range of 3-5
193
0
25
50
75
100
Al
extr
act
ed [
wt
]
Treatment time [h]
1 g H-Beta in 20 ml NH3 solution
2 g H-Beta in 20 ml NH3 solution
3 g H-Beta in 20 ml NH3 solution
0 4 6 20
Figure 54 The extent of dealumination of H-Beta zeolite with 13 M HNO3 at 100 degC at
different ratios of zeolite to HNO3 solution as a function of treatment time
The results obtained are in accordance with the observation made by Baran et al (2012)
who achieved similar degree of Al extraction from H-Beta zeolite (SiAl = 17) within 2 h and
~100 Al extraction in 4 h using the same concentration of HNO3 (the zeolite to acid solution
ratio not specified) but at 80 degC and under vigorous stirring possibly explaining faster Al
extraction rate and higher degree of dealumination The XRF results also indicate a decrease
in Si content of the dealuminated Beta zeolites This is noticeable on the samples with similar
percentage of Al extracted at relatively low SiAl ratio indicating that Si was possibly extracted
too eg deAl-BetaI6h vs deAl-BetaI20h A decrease in the Si content in conjunction with a
decrease in the BET-N2 surface area of the deAl-Beta zeolites suggest that the BEA structure
may be partially destructed The BET-N2 adsorptiondesorption isotherms of all deAl-Beta
zeolites in this work resemble the type IV adsorption isotherm according to the classification
194
by the IUPAC Commission on Colloid and Surface Chemistry (Balbuena and Gubbins 1993)
See Chapter 512 for a description of the type IV adsorption isotherm
5322 Surface acidity and the degree of Al extraction
TPD-NH3 profiles of H-Beta zeolite before and after the dealumination under the
conditions studied are shown in Figure 55 Almost complete disappearance of the NH3
desorption peaks in the TPD-NH3 profiles of deAl-Beta samples dealuminated from 5 min to
20 h as compared to their parent H-Beta zeolite indicates a significant decrease in acidity
associated with successful removal of Al atoms from the BEA framework The amount of NH3
desorbed from deAl-Beta samples dealuminated for 5 min is significantly lower as compared
to H-Beta zeolite while the profiles of the samples dealuminated for 15 min 4 6 and 20 h are
nearly identical with the amount of NH3 desorbed approaching zero
195
100 200 300 400 500 600
0
50
100
150
200
250
300
deAl-BetaI20h
deAl-BetaII6h
deAl-BetaII4h
deAl-BetaII 15 min
deAl-BetaII 5min
H-Beta zeolite
Con
cen
trati
on
of
NH
3 i
n g
as
flow
[p
pm
]
Temperature [C]
Figure 55 TPD-NH3 profiles of H-Beta zeolite before and after the treatment with 13 M HNO3
at 100 degC deAl-Beta refers to dealuminated Beta zeolite the Roman and Arabic numerals
indicate the number of grams of H-Beta zeolite dealuminated in 20 ml of 13 M HNO3 and the
dealumination time in hours or minutes respectively The data presented is from a single
measurement
The XRF analysis results show that the extraction of Al from the H-Beta zeolite
dealuminated for 4-20 h reached 97-98 wt The drop in zeolite Al content determined by
XRF correlates with the decrease of acidity determined by TPD-NH3 for the samples
dealuminated for 4-20 h However only the TPD-NH3 data is available for the samples
dealuminated for 5 and 15 min as opposed to the XRF and TPD-NH3 available for other
samples Based on the correlation of data from XRF and TPD-NH3 established on samples
dealuminated for 4-20 h it is likely that significant degree of dealumination is equally achieved
196
at the reaction times le 15 min Using identical dealumination conditions Baran et al (2012)
reported 94 Al extracted from Beta zeolite providing the SiAl ratio (bulk) of 297 after 5 min
After 2 h the degree of Al extraction reached 98 and the SiAl molar ratio (bulk) increased
from 297 to 947 The high dealumination degree but accompanied by a slow increase in the
SiAl ratio in bulk suggests that the dealumination process comprises two subsequent stages
In the first stage according to Kooyman et al (1997) Al is removed from its tetrahedral
framework position and becomes octahedrally coordinated inside the zeolitic pore structure
resulting in a significant change in the framework SiAl ratio but not the bulk SiAl ratio In
other words Al may have been extracted from the framework but is still present inside the
pores of the zeolite In the second stage the octahedrally coordinated 119860119897(1198672119874)63+ species
diffuse out of the zeolitic pore system into the solution Thus the bulk SiAl ratio decreases
whereas the framework SiAl ratio does not change The SiAl ratio of the samples would
increase after washing out the dissolved Al Therefore the SiAl ratio of dealuminated zeolite
samples may be different depending on the characterisation techniques and the data
presentation The bulk SiAl ratio can be obtained by AAS or XRF spectroscopy while the
framework SiAl ratio may be obtained by MAS NMR spectroscopy (Muumlller et al 2000)
533 Conclusions and discussion
The experiments carried out on the dealumination of H-ZSM-5 under the conditions
studied did not lead to significant Al extraction (~6 wt) On the other hand nearly complete
dealumination (97-98wt) was achieved for H-Beta zeolite samples under identical
experimental conditions ie 1 g zeolite in 20 ml of 13 M HNO3 100 degC 20 h The degree of
Al extraction was found to be much higher for Beta zeolite than ZSM-5 which is consistent
with the observation made by Gonzalez et al (2011) who concluded that the extent of
197
dealumination was a function of the zeolite structure the degree of dealumination decreased in
the following order BEA gt MOR gt MFI
Based on the experimental results the following observation were made
While 97-98 wt Al extraction was achieved in Beta zeolite an insignificant
degree of dealumination for ZSM-5 was attained possibly due to the higher
internal strains within its structure (Cha et al 2013)
The dealumination did not significantly affect the BEA structure of Beta zeolite
Its flexible structure allows it to remain intact without Al
Optimum conditions for dealumination of Beta zeolite were established 2 g
zeolite in 20 ml of 13 M HNO3 solution 100 degC 4 h
A significant decrease in the density of the acid sites of Beta zeolite was
observed even after 5 min of dealumination
54 Metal-exchanged dealuminated Beta zeolites
The deAl-BetaII4h was selected as a starting material for SSIE with the desired metal
ions ie Ce3+ La3+ Sn2+ and Zn2+ due to its high level of dealumination and the highest
relative crystallinity see Table 34 indicating its structure was not adversely affected by the
acid treatment The preparation procedure used is described in Chapter 443 The target metal
contents studied and the analytical techniques used for characterisation are listed in Table 35
The zeolites were doped with various amounts of metals in order to evaluate the effect on the
hydrothermal conversion of glycerol
198
Table 35 The list of metal-exchanged deAl-BetaII4ha prepared by SSIE and the analytical
techniques used for characterisation The metal contents are given in weight relative to the
deAl-BetaII4h
Sample Mole ratio of
metalAlextracted
Weight ratio of
metalzeolite
XRD TPD-NH3 TEM-EDX
NH4-Beta - -
H-Beta - -
deAl-BetaII4h - -
05 CedeAl-BetaII4h 004 05100
07 CedeAl-BetaII4h 006 07100
1 CedeAl-BetaII4h 009 11100
6 CedeAl-BetaII4h 057 65100
10 CedeAl-BetaII4hb 099 115100
10 LadeAl-BetaII4hb 100 115100
88 SndeAl-BetaII4hb 099 97100
5 ZndeAl-BetaII4hb 099 54100
48 CedeAl-BetaII4h (Cl)c 043 5100
48 SndeAl-BetaII4h (Cl)c 052 5100
48 ZndeAl-BetaII4h (Cl)c 094 5100
a deAl-BetaII4h refers to dealuminated Beta zeolite prepared from H-Beta zeolite dealuminated
in 20 ml of 13 M HNO3 at 100 degC for 4 h b The amount of metal equals the number of mole
of Al extracted c The metal precursors used were corresponding chlorides salts instead of
nitrates used for Ce- and La-doped or acetates for Sn- and Zn-doped catalysts
541 Physical properties
Phase identification
A series of deAl-BetaII4h catalysts containing various amounts of Ce was prepared and
characterised by XRD Figure 56 shows the XRD spectra of H-Beta zeolite and deAl-BetaII4h
compared to those of 05 wt- 07 wt- 1 wt- 6 wt- and 10 wt CedeAl-BetaII4h
199
catalysts The XRD spectra reveal that the BEA structure was preserved in all samples despite
the changes in the peak intensities d302 positions and d302-spacings summarised in Table 37
(page 207) The CeO2 was detected in the 6 wt- and 10 wt CedeAl-BetaII4h catalysts
which also developed a characteristic pale yellow discoloration of CeO2 This observation is in
good agreement with the work of Xue et al (2006) who reported the presence of CeO2 in Ce-
loaded NH4-Y zeolites at CeNH4 mole ratio which is equivalent to the CeAl mole ratio as
used in current work larger than 095 The CeO2 is known to form by thermal decomposition
of Ce(NO3)36H2O (Kamruddin et al 2004)
10 20 30 40 50 60 70
0
2000
4000
6000
8000
10000
12000
6 8 10 20 22 24
0
2000
4000
6000
8000
10000
X
XXX
H-Beta zeolite
deAl-BeatII4h
10 CedeAl-BetaII4h
6 CedeAl-BetaII4h
1 CedeAl-BetaII4h
07 wt CedeAl-BetaII4h
045 wt CedeAl-BetaII4h
Inte
nsi
ty [
cou
nts
]
2 [degree]
X
CeO2
Inte
nsi
ty [
cou
nts
]
2 (degree)
045 to 1 CedeAl-BetaII4
Figure 56 The XRD spectra of H-Beta zeolite compared to those of deAl-BetaII4h and Ce-
incorporated deAl-BetaII4h with the Ce content of 05 wt 07 wt 1 wt 6 wt and 10
wt corresponding to the CeAlextracted molar ratios of 004 006 009 057 and 099
respectively The inset graph details the two main peaks at around 2θ = 78deg (d101) and 2θ =
226deg (d302) Symbol x = CeO2
200
The presence of CeO2 in the 6 wt- and 10 wt CedeAl-BetaII4h catalysts indicates
that the incorporation of Ce into the zeolite structure either exceeded its maximum or was
unsuccessful The TEM micrograph of 10 wt CedeAl-BetaII4h shows significant amount of
CeO2 particles (dark spots) dispersed over the surface of Beta zeolite see Figure 57 The CeO2
was not as expected detected in 05 wt- 07 wt- 1 wt CedeAl-BetaII4h catalysts likely
because it was below the detection limit of XRD spectroscopy (typically ~3wt for a two
phase mixture (Toney 1992) and crystalline grain size roughly about 30Aring (Malin 2015))
Figure 57 The TEM micrograph of 10 wt CedeAl-BetaII4h ie the molar ratio of
CeAlextracted ~1 The dark spots are assumed to be CeO2
The presence of metal oxides was observed not only in CedeAl-BetaII4h with high
metal ion loadings ie the molar ratios of CeAlextracted ~057 and ~1 but also in Sn- and
ZndeAl-BetaII4h Figure 58 shows the XRD spectra of deAl-BetaII4h before and after the
SSIE with salts of various metals ie Ce3+ La3+ Sn2+ and Zn2+ at the molar ratio of metal ion
201
to Alextracted of ~1 corresponding to 10 wt Ce- 10 wt La- 88 wt Sn- and 5 wt ZndeAl-
BetaII4h The XRD patterns of all samples except for 10 wt LadeAl-BetaII4h show not
only the diffraction peaks of the parent BEA structure but also the traces of corresponding
metal oxides ie ZnO SnO2 and CeO2 see Table 36 for a summary of phases identified The
formation of metal oxides was also observed in 48 wt metal-exchanged samples using H-
and NH4-forms of both Beta zeolite and ZSM-5 as starting materials see Chapter 52 La2O3
was not detected in 10 wt LadeAl-BetaII4h which may be possibly explained by incomplete
thermal decomposition of La(NO3)3middotH2O to poorly-crystalline La3O4NO3 at 550 degC (Gobichon
et al 1996 Mentus et al 2007)
202
0
3000
6000
0
3000
6000
0
3000
6000
0
3000
6000
10 20 30 40 50 60 70
0
3000
6000
+ = NaCl
= ZnO
(D)
5 ZndeAl-BetaII4h
2 [degree]
+
+ = NaCl
O = SnO2
(C)
ooo+
88 SndeAl-BetaII4h
Inte
nsi
ty [
Co
un
ts]
+
+
+
o
+ = NaCl
(B)
++
+
10 LadeAl-BetaII4h
+ = NaCl
x = CeO2
xxx +
10 CedeAl-BetaII4h
+x
(A)
+ = NaCl
(E)
++
+
deAl-BetaII4h
Figure 58 XRD spectra of deAl-Beta zeolite before and after SSIE with salts of Ce3+ La3+
Sn2+ and Zn2+ The initial molar ratio of metal ion to Alextracted in all samples is ~1 corresponding
to 10 wt Ce- 10 wt La- 88 wt Sn- and 5 wt ZndeAl-BetaII4h NaCl used as an
internal standard was added to the samples prior to analysis Symbol + = NaCl = ZnO
ο = SnO2 x = CeO2
203
Table 36 A summary of phases identified in the XRD spectra of various metal-incorporated
deAl-BetaII4ha catalysts
Sample Mole ratio of
metalAlextracted
Phases identified
NH4-Beta - Beta zeolite
H-Beta - Beta zeolite
deAl-BetaII4h - Beta zeolite
05 wt CedeAl-BetaII4h 004 Beta zeolite
07 wt CedeAl-BetaII4h 006 Beta zeolite
1 wt CedeAl-BetaII4h 009 Beta zeolite
6 wt CedeAl-BetaII4h 057 Beta zeolite CeO2
10 wt CedeAl-BetaII4hb 099 Beta zeolite CeO2
10 wt LadeAl-BetaII4hb 100 Beta zeolite
88 wt SndeAl-BetaII4hb 099 Beta zeolite SnO2
5 wt ZndeAl-BetaII4hb 099 Beta zeolite ZnO
48 wt CedeAl-BetaII4h (Cl)c 043 Beta zeolite CeO2
48 wt SndeAl-BetaII4h (Cl)c 052 Beta zeolite SnO2
48 wt ZndeAl-BetaII4h (Cl)c 094 Beta zeolite 2 unidentified peaks a deAl-BetaII4h refers to dealuminated Beta zeolite prepared from H-Beta zeolite dealuminated
in 20 ml of 13 M HNO3 at 100 degC for 4 h b The amount of metal equals the number of mole
of Al extracted c The metal precursors used were corresponding metal chlorides instead of
nitrates used for Ce- and La-doped or acetates for Sn- and Zn-doped catalysts
The Si-O(H)-Al groups (bridging hydroxyl groups) in zeolite structure are relatively
highly acidic (Hunger 1997 Kazansky et al 2003) allowing a rapid ion exchange with a metal
cation On the other hand the Si-OH (silanol) groups formed inside the zeolite framework
after the Al removal see Equation 19 are only slightly acidic (Kazansky et al 2003) possibly
resulting in slow metal exchange rate The thermal decomposition of metal salts to metal oxides
may then be considered as a competing reaction The use of strong acid cerium salt such as
CeCl3 instead of weaker acid cerium salt such as Ce(NO3)3 may promote the ion-exchange
204
due to the higher acidity of the salt To examine the effect of chloride-based salts 48 wt
Ce- Sn- and ZndeAl-BetaII4h catalysts were prepared using corresponding metal chloride as
a precursor The XRD spectra of the three catalysts see Figure 59 reveal the presence of metal
oxides in 48 wt Ce- and 48 wt SndeAl-BetaII4h samples despite lower metal ion loadings
as compared to those prepared from metal nitrate and acetate ie 10 wt Ce- and 88 wt Sn-
deAl-BetaII4h respectively On the other hand ZnO was not detected in 48 wt ZndeAl-
BetaII4h prepared from ZnCl2 although the Zn2+ loading was similar to that prepared from
Zn(acetate)2 ie 5 wt Zn2+ relative to zeolite and two unidentified peaks at 2θ = 565deg and
691deg were detected in its XRD spectra The results indicate that the nature of metal salt
precursors does not affect the extent of metal oxide formation
205
0
3000
6000
9000
0
3000
6000
0
3000
6000
10 20 30 40 50 60 70
0
3000
6000
5 wt ZndeAl-BetaII4h
48 wt ZndeAl-BetaII4h (Cl)
2 [degree]
+
(D)
(B)
(C)
88 wt SndeAl-BetaII4h
48 wt SndeAl-BetaII4h (Cl)
ooo+
Inte
nsi
ty [
Cou
nts
]
+
+
+
o
10 wt CedeAl-BetaII4h
48 wt CedeAl-BetaII4h (Cl)
+
x
x + x x
(A)
++
+
deAl-BetaII4h
Figure 59 XRD spectra of (D) deAl-BetaII4h before and after SSIE with salts of (A) Ce3+ (B)
Sn2+ and (C) Zn2+ The black colour denotes catalysts prepared using the nitrate or acetate-
based salts while the blue colour denotes the sample prepared using the metal chloride In the
former the molar ratio of metal ion to Alextracted in all samples is ~1 NaCl used as an internal
standard was added to the samples prior to analysis Symbol + = NaCl = ZnO ο = SnO2 x
= CeO2
206
Changes in XRD spectra
Independently of the type of metal and the metal content the BEA structure of all
catalysts prepared was preserved after the SSIE however the changes in the peak intensities
d302 position and d302-spacing were noted and tabulated See Figure 60 below Figure 121 and
122 in the Appendix for the XRD spectra of 10 wt La- 88 wt Sn- and 5 wt ZndeAl-
BetaII4h catalysts respectively compared to those of H-Beta zeolite and deAl-BetaII4h Table
37 summarises the shifts of the main diffraction peaks (d302) the changes of d-spacing (d302)
and the relative crystallinities of deAl-BetaII4h before and after the SSIE reaction with metal
salts under study
10 20 30 40 50 60 70
0
2000
4000
6000
8000
10000
12000
6 8 10 20 22 24
0
2000
4000
6000
++
+ NaCl
H-Beta zeolite
deAl-BetaII4h
10wt LadeAl-BetaII4h
Inte
nsi
ty [
co
un
ts]
2 [degree]
+
Inte
nsi
ty [
cou
nts
]
2 [degree]
Figure 60 The XRD spectra of H-Beta zeolite compared to those of deAl-BetaII4h and 10
LadeAl-BetaII4h The inset graph details the two main peaks at around 2θ = 78deg (d101) and 2θ
= 226deg (d302) NaCl used as an internal standard at the zeolite to NaCl mass ratio of 267 was
added to the samples prior to analysis Symbol + = NaCl
207
Table 37 Summary of the main diffraction peaks positions d-spacing (d302) (2θ = 225ordm) and
the relative crystallinity of NH4-Beta and H-Beta zeolites as well as deAl-BetaII4h zeolite
before and after the SSIE reaction with various metal salts
Sample Δd302
position
[degree]a
Crystallite
sizeb
[nm]
d302-
spacing
[Aring]
Δd302 [Aring]c
Relative
crystallinityd
NH4-Beta +009 159 394 100
H-Beta 000 151 396 000 105
deAl-BetaII4he +019 129 392 -004 100
045 CedeAl-BetaII4h +022 118 392 -004 82
07 CedeAl-BetaII4h +019 131 392 -004 79
1 CedeAl-BetaII4h +019 117 392 -004 78
6 CedeAl-BetaII4h +012 127 393 -002 64
10 CedeAl-BetaII4hf +038 110 389 -007 58
10 LadeAl-BetaII4h +016 110 393 -003 63
88 SndeAl-BetaII4h +012 135 393 -002 68
5 ZndeAl-BetaII4h +003 146 395 -001 92
48 CedeAl-BetaII4h (Cl)g +012 127 393 -002 64
48 SndeAl-BetaII4h (Cl)g +012 135 393 -002 68
48 ZndeAl-BetaII4h (Cl)g +003 146 395 -001 92
a Changes of the position of the main diffraction peak 2θ = 2247deg (d302 ) relative to H-Beta zeolite the
ldquo+rdquo (plus symbol) denotes the positive peak shift relative to parent H-Beta zeolite b The primary
crystalline size was calculated by Scherrerrsquos equation c Changes of the d-spacing (d302) relative to H-
Beta zeolite the ldquo-rdquo (minus symbol) denotes a negative d-spacing change d The crystallinity of the
samples relative to NH4-Beta zeolite based on the intensity of its main diffraction peak (2θ = 226deg) in
the XRD spectra e deAl-BetaII4h was prepared using 2 g of H-Beta zeolite in 20 ml of 13 M HNO3
100 degC 4 h f The amount of Ce introduced equals the number of mol of Alextracted g corresponding metal
chloride was used as a precursor The error is in the range of 3-5
208
The SSIE of deAl-BetaII4h with salts of Ce3+ La3+ Sn2+ and Zn2+ resulted in
(i) a decrease in the relative crystallinity of parent zeolite
(ii) significant changes in the position of the main diffraction peak at 2θ = 226deg
(d302) and
(iii) changes in d302-spacing (also known as interplanar distance)
The incorporation of a cation may induce significant changes of the diffracted peak
intensities due to different X-ray radiation absorption coefficient of individual ion (Medrud
1999) The changes of the position of the d302 peak at 2θ = 226deg are associated with the changes
in the interplanar distance (d302-spacing) (see Braggrsquos law Equation 9 in Chapter 413) within
the crystal lattice indicating a contracting or expanding BEA framework For instance a
decrease from 3956 Aring (H-Beta zeolite with 2θ of 22470deg) to 3921 Aring (deAl-BetaII4h with
2θ of 22662deg) is consistent with the removal of aluminium atoms while an increase from 3921
Aring (deAl-BetaII4h) to 3948 Aring at 22501 (5 wt ZndeAl-BetaII4h) indicates the incorporation
of Zn ions inside the framework may be successful (Dzwigaj et al 2013) However the XRD
patterns of 5 wt ZndeAl-BetaII4h also reveal the presence of ZnO indicating that Zn at least
partially co-exists as a separate phase
To conclude whether or not the incorporation of metal ions into the BEA framework
was successful a combination of several characterisation techniques such as pyridine
adsorption by FTIR and MAS NMR would be required (Kazansky et al 2003 Tang et al
2014) Both techniques can be used to determine the existence and concentration of Lewis acid
sites associated with the presence of metals as well as to identify the environment of each
element in each preparation steps For instance the presence of silanol groups in the
dealuminated zeolite before and after the SSIE see Equation 19 (page 133)
209
542 Surface acidity
Figure 61 shows the TPD-NH3 profiles of the H-Beta zeolite before and after the
treatment with 13 M HNO3 at 100 degC for 4 h (deAl-BetaII4h) compared to those of 10 wt
Ce- 10 wt La- and 88 wt Sn-incorporated deAl-BetaII4h samples see Chapter 513 and
532 for a full discussion on the surface acidity of H-Beta zeolite and deAl-Beta zeolites
respectively The TPD-NH3 profiles show that the dealumination of H-Beta zeolite resulted in
a complete disappearance of its NH3 desorption peaks indicating a successful removal of Al
atoms from the BEA framework (see the TPD-NH3 curves of H-Beta zeolite vs deAl-BetaII4h)
After the SSIE the TPD-NH3 profile of the deAl-BetaII4h shows extensively overlapping
peaks resulting in a long plateau from (i) 150 to 250 degC (ii) 125 to 200 degC and (iii) 150 to 200
degC for Ce- La- and Sn- incorporated deAl-BetaII4h samples respectively This may be due to
the metals creating new Lewis acid sites (Jentys et al 1997)
210
100 200 300 400 500 600
0
50
100
150
200
250
300
88 wt SndeAl-BetaII4h
10 wt LadeAl-BetaII4h
10 wt CedeAl-BetaII4h
deAl-BetaII4h
H-Beta zeolite
Con
cen
trati
on
of
NH
3 i
n g
as
flow
[p
pm
]
Temperature [C]
Figure 61 TPD-NH3 profiles of H-Beta zeolite before and after the treatment with 13 M HNO3
2 g zeolite in 20 ml HNO3 (aq) 100 degC 4 h (deAl-BetaII4h) compared to those of metal-
incorporated deAl-BetaII4h The percentages reported are in wt indicating the metal content
relative to zeolite
55 Immobilisation and agglomeration of Beta zeolite-
based catalysts
Zeolite powders are unsuitable for a direct use in a continuous flow reactor due to the
tendency to get washed out It is therefore necessary to shape the powder into a macroscopic
form or immobilise it on a larger particles prior to administration Several techniques were
used to process the powders into a more suitable form
211
Ceramic hollow fibre membranes (CHFM) were used to support the CeH-Beta zeolite
for its high temperature resistance and chemical mechanical stability The use of CHFM may
also alleviate reactor pressure drop usually found in packed bed reactors However only
insignificant amounts of catalysts were successfully deposited into the pores of the membrane
and the zeolite particles were washed out from the continuous flow reactors after a few hours
use See Chapter 551
The incorporation of Beta zeolite-based catalysts into the pores of mesoporous silica
matrix TUD-1 was then carried out The pellets of the Beta zeolite-TUD-1 composite
disintegrated and leached out from the reactor See Chapter 552
The extrusion of Beta zeolite-based catalysts was performed using bentonite clay binder
but the extruded catalysts partially disintegrated The Beta zeolite-based catalysts was then
extruded with γ-Al2O3 binder and methyl cellulose plasticiser in order to improve the
consistency of the extruded paste See Chapter 553
551 CeH-Beta zeolite on ceramic hollow fibre membranes
The Ce-exchanged H-Beta zeolite (CeH-Beta) was introduced into the pores of the
ceramic hollow fibre membranes (CHFM) according to the 4-step procedure described in
Chapter 444 In the final step the CeH-Beta recovered from the filtrate as well as the CeH-
Beta deposited on the CHFM were characterized by SEM-EDS and XRD
Figure 62 shows the XRD spectra of (A) H-Beta zeolite prepared by calcining of
commercial NH4-Beta at 550 degC 5 h (B) synthesized CeH-Beta powder recovered from the
filtrate and (C) CeH-Beta deposited on CHFM support and the CHFM support The XRD
pattern of the solid recovered from the filtrate collected after the synthesis confirmed the
212
presence of Beta zeolite see Figure 62A and B However the XRD spectra of the CeH-Beta
zeolite on CHFM support shows only the diffraction peaks of Al2O3 the main component of
the CHFM see Figure 62C possibly indicating that the amount of CeH-Beta zeolite
introduced is below the detection limit ie lt3wt
213
0
1000
2000
0
1000
2000
10 20 30 40 50 60 70
0
1000
2000
Inte
nsi
ty [
Co
un
ts]
(B) Synthesized zeolite (powder)
2 [degree]
(A) H-Beta zeolite
2= 7776
2= 22630
2= 25456
2= 27158
2= 43971 2= 55980
2 = 7643
2= 22557
2 = 25356
2 = 26958
2 = 439670
(C) CHFM support
after the zeolite synthesis
before the zeolite synthesis
Figure 62 XRD spectra of (A) commercial H-Beta zeolite (CP814E SiO2Al2O3 = 25) (B)
CeH-Beta recovered from the filtrate and (C) ceramic hollow fibre membrane (CHFM)
support and CeH-beta deposited on CHFM support
214
The SEM micrographs of CeH-Beta particles deposited on the CHFM are shown in
Figure 63 The zeolite particles appear to reside inside the CHFM pores
Figure 63 SEM micrographs of CeH-Beta particles on the ceramic hollow fibre membranes
The effect of the catalyst on the hydrothermal oxidation of glycerol was studied in a
continuous flow reactor according to the experimental procedures in Chapter 462 The reactor
was packed with 24 g CeH-Beta on CHFM and fitted with metal mesh at both ends The
reaction was performed at 175 degC 35 bar and a residence time of 180 and 240 s using 02 M
glycerol with the molar glycerolH2O2 ratio = 1 The results are shown in Table 38 A moderate
degree of conversion (23-25 mol) and high selectivity to liquid products (54-103 mol) were
obtained Several liquid products were detected Formic acid the main product was obtained
with the highest selectivity of 542 mol and yield of 134 mol
215
Table 38 The products of hydrothermal oxidation of 02 M glycerol with H2O2 using 24 g CeH-Beta on ceramic hollow fibre membranes as
catalyst Reaction conditions 175 degC 35 bar and a residence time of 180 and 240 s the molar glycerolH2O2 = 1
Temperature
[degC]
Residence time
[s]
Conversion
[mol]
Selectivity to liquid products [C-mol] Total
[C-mol] AA LA FA DHA GCA PVA
175 180 251 plusmn 12 153 plusmn 11 50 plusmn 07 542 plusmn 16 131 plusmn 11 114 plusmn 08 32 plusmn 04 1025 plusmn 35
175 240 232 plusmn 31 121 plusmn 16 29 plusmn 05 276 plusmn 37 68 plusmn 09 48 plusmn 06 24 plusmn 03 542 plusmn 72
Yield of liquid products [C-mol] Total
[C-mol] AA LA FA DHA GCA PVA
175 180 251 plusmn 12 38 plusmn 01 13 plusmn 02 136 plusmn 02 33 plusmn 03 29 plusmn 02 08 plusmn 01 257 plusmn 01
175 240 232 plusmn 31 28 plusmn 01 07 plusmn 01 64 plusmn 03 16 plusmn 01 11 plusmn 06 06 plusmn 03 125 plusmn 05
Note AA= acetic acid LA = lactic acid FA = formic acid DHA = dihydroxyacetone GCA = glycolic acid and PVA = pyruvaldehyde The
numbers presented are an averaged plusmn standard error of means (119878119864)
216
Despite the promising results the CeH-Beta particles deposited on the CHFM were
unsuitable for use in the continuous flow reactor as the zeolites particles washed out after 4 h
see Figure 64 Further work was abandoned
Figure 64 SEM micrograph of the CeH-Beta on CHFM after the 4 h use in the hydrothermal
oxidation of 02 M glycerol with H2O2 at 175 degC 35 bar and a residence time of 180 s the
molar glycerolH2O2 of 1 and 24 g of the catalyst
552 H-Beta and metal-dopedH-Beta zeolites embedded in TUD-1
matrix
Mesoporous silica matrix TUD-1 was used for the agglomeration of Beta zeolite
catalysts for its easy preparation and low cost (Jansen et al 2001) The key characteristics of
this matrix include well-defined pores high thermal and hydrothermal stability (Jansen et al
217
2001) It also has high surface area and three-dimensional porosity which may be beneficial
for internal mass transfer (Lima et al 2010)
Three zeolite-based catalysts ie H-Beta 48 CeH-Beta and 48 CuH-Beta were
prepared by SSIE and incorporated into the pores of TUD-1 according to the procedure
described in Chapter 445 Cu is known as an effective catalyst for oxidation and hence it was
also used The agglomerated catalysts were characterised by XRD and BET-N2 analysis
Figure 65 shows the XRD spectra of (B) H-Beta (C) 48 wt CeH-Beta and (D) 48
wt CuH-Beta before and after incorporating into the TUD-1 matrix at the zeoliteTUD-1
weight ratio set constant to 4060 The XRD patterns of the H-Beta and the H-Beta in TUD-1
(Figure 65A amp B) do now show significant differences with the characteristic diffraction peaks
of BEA structure of Beta zeolite as the single phase identified The apparent broadening of low
angle peaks may be explained by an intergrowth of different polymorphs of the Beta zeolite
(mainly A and B) in the crystals (Lima et al 2010) The apparent decrease in intensity of
agglomerated H-Beta in TUD-1 matrix may be explained by lower zeolite content relative to
pure H-Beta zeolite see Figure 65B The characteristic diffraction peaks of Beta zeolite as well
as the low angle peaks broadening are also observed for the 48 wt CeH-Beta and 48 wt
CeH-Beta in TUD-1 with CeO2 as an additional phase present detected in both spectra On the
other hand no line broadening was observed for the 48 wt CuH-Beta in TUD-1 (Figure
65D) with Beta zeolite and CuO being the only phases identified
The formation of amorphous silica matrix TUD-1 may be confirmed by XRD analysis
but only at lower scanning angles (1deg to 5deg) (Antunes et al 2012)
218
10 20 30 40 50 60 70
0
750
1500
2250
0
750
1500
2250
0
750
1500
2250
0
750
1500
2250
(D)
= CuO
Inte
nsi
ty [
Cou
nts
]
2 [degree]
48 CuH-Beta in TUD-1
48 wt CuH-Beta
(C)
xxx
x = CeO2
48 CeH-Beta in TUD-1
48 CeH-Beta
x
(B)
H-Beta in TUD-1
H-Beta
(A)
H-Beta
Figure 65 The XRD spectra of Beta zeolite-based catalysts in mesoporous silica matrix TUD-
1 (B) H-Beta in TUD-1 (C) 48 wt CeH-Beta in TUD-1 and (D) 48 wt CuH-Beta in
TUD-1 The XRD spectra of all the catalysts before incorporating into the TUD-1 are given for
comparison The Beta zeolite content is set constant to 40 wt zeolite relative to TUD-1
Symbols x = CeO2 = CuO
219
The BET-N2 adsorption and desorption isotherms of (i) H-Beta vs H-Beta in TUD-1
and (ii) H-Beta 48 wt CuH-Beta and 48 wt CuH-Beta in TUD-1 are shown in Figure
123 and 124 (in the Appendix) respectively Similar to H-Beta zeolite (see Figure 42 page
162) the BET-N2 adsorptiondesorption isotherms resemble the type IV adsorption isotherm
which is typical for mesoporous materials (Balbuena and Gubbins 1993) See Chapter 512
for a description of the type IV adsorption isotherm
The incorporation of the H-Beta zeolite and 48 wt CuH-Beta zeolite catalysts into
the TUD-1 matrix leads to an increase in their BET-N2 surface areas see Table 39 An increase
in the surface area of the TUD-1 incorporated samples may be due to multilayer adsorption on
the external surface of the nanocrystallites
Table 39 BET-N2 analysis results of H-Beta zeolite and 48 wt CuH-Beta zeolite before and
after the incorporation into the mesoporous silica matrix TUD-1 The data presented is from a
single measurement
Sample Micropore
volume
[cmsup3middotg-1]
Micropore
area
[msup2middotg-1]
External
surface area
[msup2middotg-1]
BET surface
area
[msup2middotg-1]
H-Beta zeolite 017 367 231 598
H-Beta zeolite in TUD-1 009 195 428 624
48 wt CuH-Beta 015 330 200 530
48 wt CuH-Beta in TUD-1 012 278 348 627
The protocol used in this work was previously reported by Lima et al (2010) to provide
the H-Beta zeolite incorporated in the TUD-1 silica matrix The characterisation techniques
220
used in this work do not allow to confirm conclusively that the incorporation of Beta zeolite-
based catalysts into TUD-1 matrix was achieved
The H-BetaTUD-1 composite material was pelletized and tested for the hydrothermal
oxidation of 02 M glycerol with H2O2 at 175 degC 35 bar 180 s using the molar glycerolH2O2
ratio of 1 The experiment procedure is provided in Chapter 462 The pellets disintegrated and
leached out from the reactor after a few hours The chemical analysis of reaction products was
not performed
553 Extruded catalysts
5531 Starting zeolite-based catalysts
Extrusion is another technique used to shape the zeolite powders into macroscopic
forms more suitable for administration in a continuous flow reactor Three zeolite-based
catalysts ie
H-Beta
25wt CuH-Beta and
48wt CuH-Beta
were prepared mixed with either bentonite or γ-Al2O3 binders according to the procedure
described in Chapter 446 and then extruded into 3-5 mm long 2 mm diameter rods The rods
were calcined at 550 degC prior to the instrumental analysis H-Beta zeolite was prepared by the
calcination of NH4-Beta zeolite (see Chapter 441) while the 25wt and 48wt CuH-Beta
zeolite catalysts were prepared by SSIE according to the procedure described in Chapter 442
221
The 25wt CuH-Beta zeolite was used as the results obtained on the 48wt CuH-Beta
zeolite indicated that lowering the Cu loading may increase the yield of liquid products
Figure 66 presents the XRD spectra of H-Beta zeolite 25 wt CuH-Beta and 48 wt
CuH-Beta The XRD pattern of H-Beta zeolite shows the characteristic peaks of the BEA
structure with the main diffraction peaks at 2θ = 75-8deg and 226deg The XRD patterns of H-Beta
zeolite after the SSIE with nitrate salt of Cu2+ ie 25 wt- and 48 wt CuH-Beta catalysts
show that the BEA structure of Beta zeolite remained preserved but the CuO was identified as
an additional phase in both samples The XRD spectrum of 25 wt CuH-Beta is almost
identical to that of 48 wt CuH-Beta but exhibits lower intensity of the diffraction peaks of
CuO
10 20 30 40 50 60 70
0
1000
2000
3000
4000
5000
6000
2 [degree]
= CuO
H-Beta
25 wt CuH-Beta
48 wt CuH-Beta
Inte
nsi
ty [
cou
nts
]
Figure 66 The XRD spectra of H-Beta zeolite 25 wt CuH-Beta zeolite and 48 wt CuH-
Beta zeolite Symbol = CuO
222
5532 Beta zeolite-based catalysts extruded with bentonite clay
I Powder X-ray diffraction
The XRD spectra of the H-Beta zeolite and 48 wt CuH-Beta zeolite catalysts before
and after the extrusion with bentonite clay are shown in Figure 67 The decreased intensities
of the extruded zeolites relative to pure zeolites may be explained by the dilution with the
bentonite as the zeolite catalyst content in the extruded form was 68 wt However the
diffraction peaks of bentonite and methylcellulose see Figure 125 in the Appendix are absent
in the XRD spectra of the extruded catalysts calcined at 550 degC which may be explained by the
poor crystallinity of bentonite and the thermal instability at 550 degC for methylcellulose
223
10 20 30 40 50 60 70
0
750
1500
2250
0
750
1500
2250
0
750
1500
2250
0
750
1500
2250
= CuO
Inte
nsi
ty [
Cou
nts
]
2 [degree]
(D) extruded 48 wt CuH-Betabentonite
(C) 48 wt CuH-Beta
(B) extruded H-Betabentonite
(A) H-Beta
Figure 67 The XRD spectra of (A) amp (B) H-Beta zeolite and (C) amp (D) 48 wt CuH-Beta
zeolite before and after the extrusion with bentonite clay The Cu content is given in weight
relative to H-Beta zeolite The content of ldquo48 wt CuH-Betardquo is 68 wt relative to bentonite
224
II BET N2 and SEM analysis
The BET N2 adsorptiondesorption isotherms of extruded H-Beta and 48 wt CuH-
Beta see Figure 126 (in the Appendix) resemble the type IV isotherms featuring the H1
hysteresis loops typical for mesostructured materials with pore sizes larger than 40 Aring (Balbuena
and Gubbins 1993 Melero et al 2008) The extrusion of H-Beta zeolite and 48 wt CuH-
Beta zeolite with bentonite resulted in a decrease of 11 in their BET surface area mainly due
to a decrease in the total pore volume and external surface area see Table 40 This may be
explained by the dilution of the mesostructured material with the low-surface area bentonite
clay The SEM micrographs of H-Beta and 48 wt CuH-Beta before and after extrusion with
bentonite see Figure 68 show that after the extrusion most zeolite particles are encapsulated
within the clay body
Table 40 BET N2 surface area analysis results for H-Beta zeolite and 48wt CuH-Beta
zeolite before and after the extrusion with clay (bentonite) The metal contents are given in
weight percentage (wt) relative to H-Beta zeolite The weight percentage of H-Beta zeolite
or ldquo48 wt CuH-Betardquo is 68wt relative to bentonite The data presented is from a single
measurement
Sample Micropore
volume
[cmsup3middotg-1]
Micropore
area
[msup2middotg-1]
External
surface area
[msup2middotg-1]
BET N2
surface area
[msup2middotg-1]
H-Beta zeolite 017 367 231 598
extruded H-Betabentonite 012 269 154 423
48 CuH-Beta 015 330 200 530
extruded 48 CuH-Betabentonite 011 247 141 388
bentonite 001a - - 94a
a reported by Grisdanurak et al (2012)
225
Figure 68 SEM micrographs of (A) H-Beta (B) H-Beta extruded with bentonite (C) 48 wt
CuH-Beta and (D) 48 wt CuH-Beta extruded with bentonite The metal contents are given
in weight percentage (wt) relative to H-Beta zeolite The weight percentage of H-Beta zeolite
or ldquo48 wt CuH-Betardquo is 68 wt relative to bentonite
5533 Beta zeolite-based catalysts extruded with γ-Al2O3
The procedure used for preparation of extruded catalysts is described in Chapter 446
All extruded catalyst samples discussed in the chapters below were calcined at 550 degC prior to
the instrumental analysis
I Powder X-ray diffraction
The XRD spectra of H-Beta zeolite 48 wt- and 25 wt CuH-Beta before and after
extrusion with γ-Al2O3 are shown in Figure 69 70 and 71 respectively The XRD patterns of
all three samples show that the BEA structure remained preserved after the extrusion with γ-
226
Al2O3 followed by calcination at 550 degC despite the reduction in the intensity of the diffraction
peaks possibly caused by the relatively lower zeolite content in the extruded samples The
reduction in the peak intensity of zeolite-based catalysts extruded with γ-Al2O3 is however
not as high as in the case of bentonite clay (see Figure 67 Chapter 5532) In addition to the
characteristic diffraction peaks of Beta zeolite the XRD patterns of the catalysts show the
presence of the following phases
(i) CuO in 25 wt- and 48 wt CuH-Beta both before and after extrusion with
γ-Al2O3 followed by calcination at 550 degC
(ii) γ-Al2O3 in all extruded catalysts see Figure 130 (in the Appendix) for the XRD
patterns of supporting extrusion materials ie γ-Al2O3 and methyl cellulose
(MC)
The addition of MC 5 wt relative to zeolite + γ-Al2O3 into the pastes of (i) H-Beta
zeolite or (ii) 48 wt CuH-Beta zeolite increased the homogeneity facilitating the material
processing The XRD spectra of both H-Betaγ-Al2O3 and 48 wt CuH-Betaγ-Al2O3
extruded with and without MC show no significant differences see Figure 69(B) vs (C) and
Figure 70(B) vs C likely due to the fact that MC decomposed during the calcination at 550
degC However a reduction in the intensity of the main diffraction peak appearing at 2θ = 226deg
(d302) was noted for the extruded catalysts prepared from the paste with MC see Table 41 for
a summary of composition of extruded catalysts
A slight shift of the main diffraction peak was also observed for all the samples after
the extrusion but the cause of the shift is uncertain as an internal standard such as NaCl was
not used
The reduction in the diffraction peak intensity of the zeolite catalysts resulted in a
decrease in their XRD relative crystallinity Table 42 For instance the relative crystallinity of
227
H-Beta zeolite after the extrusion with γ-Al2O3 with and without MC decreases from 105 to
49 and 38 with respect to NH4-Beta The changes in the intensity and position of the main
diffraction peaks as well as the BET N2 analysis results are summarised in Table 42
10 20 30 40 50 60 70
red circle-Al2O
3
(C) extruded H-Beta-Al2O
3
(B) extruded H-Beta-Al2O
3(MC)
(A) H-Beta
Inte
nsi
ty [
arb
un
its]
2 [degree]
Figure 69 The XRD spectra of H-Beta zeolite before and after the extrusion with γ-Al2O3
binder with or without methyl cellulose (MC) acting as a plasticiser The weight percentage of
zeolite relative to binder in the extrusion paste was 60 (without MC) and 575 (with MC)
Symbol red circle = γ-Al2O3
228
10 20 30 40 50 60 70
= CuO red circle-Al2O
3
(C) extruded 48 wt CuH-Beta-Al2O
3
(B) extruded 48 wt CuH-Beta-Al2O
3(MC)
Inte
nsi
ty [
arb
un
its]
2 [degree]
(A) 48 wt CuH-Beta
Figure 70 The XRD spectra of 48 wt CuH-Beta zeolite before and after the extrusion with
γ-Al2O3 binder with or without methyl cellulose (MC) plasticiser The metal content is 48 wt
Cu relative to H-Beta zeolite The composition of the extrusion paste was 575 wt zeolite
catalyst 375 wt γ-Al2O3 and 5wt MC (B) and 60 wt zeolite catalyst 40 wt γ-Al2O3
(C) Symbol = CuO red circle = γ-Al2O3
229
10 20 30 40 50 60 70
= CuO red circle-Al2O
3
(A) extruded 25 wt CuH-Beta-Al2O
3 (MC)
Inte
nsi
ty [
arb
un
its]
2 [degree]
(B) 25 wt CuH-Beta
Figure 71 The XRD spectra of 25 wt CuH-Beta zeolite before and after extrusion with γ-
Al2O3 binder using methyl cellulose (MC) plasticiser The metal content is 25 wt Cu relative
to H-Beta zeolite The composition of the extrusion paste was 575 wt zeolite catalyst 375
wt γ-Al2O3 and 5wt MC Symbol = CuO red circle = γ-Al2O3
Table 41 Composition of extrusion paste for the extrusion of H-Beta zeolite and 48 wt
CuH-Beta with γ-Al2O3 The metal content is 48 wt Cu relative to H-Beta zeolite
Catalysts
Composition [wt] Normalized
wt of zeolite
catalyst relative
to γ-Al2O3
H-Beta or
48 CuH-Beta γ-Al2O3 MC
extruded H-Betaγ-Al2O3 60 40 - 6040
extruded H-Betaγ-Al2O3 (MC) 575 375 5 605395
extruded 48 CuH-Beta γ-Al2O3 60 400 - 6040
extruded 48 CuH-Beta γ-Al2O3 (MC) 575 375 5 605395
230
Table 42 Physical properties of Beta zeolite-based catalysts extruded with bentonite clay or γ-
Al2O3 The weight percentage assigned in CuH-Beta indicates the metal content relative to H-
Beta zeolite The weight percentage of H-Beta and CuH-Beta catalysts relative to bentonite in
the extrusion paste is 60 (without MC) and 575 (with MC) Note MC = methyl cellulose
Catalyst BET N2
surface area
[m2middotg-1]
Micropore
Volume
[cmsup3middotg-1]
XRD relative
crystallinitya
[]
Post of d302
[degree 2θ]
TEM-EDS
SiO2Al2O3
mol ratio
NH4-Beta 575 017 100 226 25b
H-Beta 598 017 105 225 25 b
25 wt CuH-Beta 573 016 nad 226
48 wt CuH-Beta 534 015 83 227
H-Betaγ-Al2O3c - - 49 226
H-Betaγ-Al2O3 (MC)c 423 010 38 228 54
25 wt CuH-Betaγ-
Al2O3 (MC)c 419 009 nad 226 24
48 wt CuH-Betaγ-
Al2O3c
- - 56 227 44
48 wt CuH-Betaγ-
Al2O3 (MC)c 446 010 62 227 45
a Relative crystallinity of the samples using the main peak intensity from XRD patterns (2θ = 22563deg d302) and considering
the crystallinity of NH4-Beta zeolite as 100 b manufacture specification c in an extruded form d not applicable as
different sample holder was used
II BET N2 and TEM-EDS analysis
Figure 127 128 and 129 (in the Appendix) show the BET N2 adsorptiondesorption
isotherms of H-Beta 25 wt CuH-Beta and 48 wt CuH-Beta zeolites before and after the
extrusion with γ-Al2O3 and calcination at 550 degC Similar to H-Beta zeolite (see Chapter 512)
the BET N2 adsorptiondesorption isotherms of extruded H-Betaγ-Al2O3 as well as 25 wt
CuH-Beta and 48 wt CuH-Beta catalysts before and after the extrusion with γ-Al2O3
resemble the type IV adsorption isotherm according to the IUPAC classification (Balbuena
231
and Gubbins 1993) featured with H1 hysteresis loops See Chapter 512 for a description of
the type IV adsorption isotherm
The 48 wt CuH-Beta catalyst prepared via SSIE reaction between H-Beta zeolite
and Cu2+ nitrate exhibits ~11 lower BET N2 surface area and micropore volume as compared
to its parent H-Beta zeolite see Table 42 The extrusion of the catalyst with γ-Al2O3 leads to a
further decrease of 16 in its surface area and 34 in the pore volume relative to 48 wt
CuH-Beta The decrease in the BET N2 surface area and micropore volume may possibly be
due to the blockage of the zeolite pores by metal aggregates andor by γ-Al2O3 Similarly H-
Beta and 25 wt CuH-Beta after the extrusion exhibit (i) 29 and 27 respectively lower
surface area and (ii) 42 and 45 respectively lower pore volume
TEM micrographs of extruded H-Betaγ-Al2O3 (MC) 25 wt CuH-Betaγ-Al2O3
(MC) and 48 wt CuH-Betaγ-Al2O3 (MC) are shown in Figure 72 while the results of TEM-
EDS analysis are provided in Table 42
Figure 72 TEM micrographs of (A) H-Betaγ-Al2O3 (MC) (B) 25 wt CuH-Betaγ-Al2O3
(MC) and (C) 48 wt CuH-Betaγ-Al2O3 (MC)
232
6 Hydrothermal dehydration of glycerol
Dehydration is one of the most intensively investigated reactions for glycerol The
results of study on hydrothermal dehydration of glycerol with and without a zeolite-based
catalyst carried out in batch reactors are presented in this chapter The details of experimental
procedures are provided in Chapter 45
61 Non catalysed hydrothermal dehydration of glycerol
An experimental study on the hydrothermal decomposition of glycerol without a
catalyst at varying temperature and reaction time was carried out The data obtained was used
as a benchmark to evaluate the activity of various zeolite-based catalysts
A summary of experimental conditions studied and the detected reaction products are
provided in Figure 73 and Table 43 In general the glycerol conversions obtained were rather
low but gradually increased with increasing temperature and reaction time giving a maximum
of ~59 mol at 330 degC 128 bar (saturated vapour pressure) after 300 min Lowering the
temperature from 330 to 300 degC 86 bar resulted in 306 mol conversion and only 53 mol
at 270 degC 55 bar Under identical conditions (ie 300 degC saturated vapour pressure 60 min)
similar glycerol conversion of 49 mol was previously reported by Watanabe et al (2007)
On the other hand at the same temperature of 300 degC a complete conversion of 089 M glycerol
solution was achieved within 40 min when the reactions were performed at 2955 bar
(Qadariyah et al 2011) This may be attributed to the higher pressure applied indicating that
at a given temperature the pressure higher than the saturated vapour pressure increases the
conversion rate
233
0 60 120 180 240 3000
20
40
60
80
100
0 60 120 180 240 3000
10
20
30
40
50
270 C
300 C
330 C
Gly
cero
l co
nv
ersi
on
[m
ol
]
Reaction time [min]
270 C
300 C
330 C
Sel
ecti
vit
y t
o a
ll l
iqu
id p
rod
uct
s [C
-mo
l]
Reaction time [min]
Figure 73 Glycerol conversion and selectivity to all liquid products as a function of reaction time obtained in the hydrothermal dehydration of 01 M glycerol
(03 mmol) at various temperatures under saturated vapour pressure (55-186 bar) The error bars show the standard error of mean (119878119864)